Systems of Linear Equations 3.1 Substitution Method for a Linear System 3.2 Systems of Linear Inequalities 3.3 Characteristics of a System of Linear Inequalities 3.4 Solving Systems of Linear Equations with More than Two Unknowns: Elimination Method The summary will review the most important ideas, not necessarily all the terms that were bolded in the body of the section.}This chapter covers a very important class of mathematical problems. Here, you will learn to find the values of two or more variables at the same time. These problems occur frequently in business and economics, as well as in the social sciences, life sciences, and physical sciences. A manufacturer, for example, may need to know how much can be produced of two different products using information about the available amounts of two different raw materials. A marketing manager may want to find out how effective her advertising is in two different magazines, based on information about the responses from two months of advertising. We’ll begin our investigation by looking at a classic example of diet design, based on consumption of only two foods, for simplicity. As we shall see in section 3.1, the information that comes with these problems will enable us to write two linear equations that describe the relationships between these variables. The two equations will comprise a system of simultaneous equations. For problems with two variables and two equations, the solution takes only a few steps. Use this section as an opportunity to practice “seeing” the linearity in the description of a problem and to write the equations that match the description. Sections 3.2 and 3.3 explore the topic of linear inequalities. When a linear equation is written with a less than (<) or greater than (>) sign rather than an equal (=) sign, the solution is not a line but an entire half-plane in the coordinate system. Linear inequalities have important applications in business problems where a decision maker is trying to find a combination of variables that maximizes profit or minimizes cost, subject to certain restrictions on the variables. Section 3.4 returns to the topic of simultaneous linear equations but considers problems that have three or more variables. The simple method of section 3.1 does not work as easily on these larger problems. Fortunately, an alternative method, the “elimination method,” does work well and can even be performed by spreadsheets and graphing calculators. But you should still focus on learning how to read an application problem and to write out the system of equations that describes it. The application exercises are good practice for this. 3.1 Substitution Method for a Linear System Learning Objectives Write a system of linear equations corresponding to a verbal description. Solve a system of linear equations graphically. Solve a system of linear equations algebraically by the substitution method. Solve applied problems using tables, graphs and systems of equations. Chapter 3 Systems of Linear Equations and Inequalities 2 Spotlight on Designing a Diet An important task in hospitals, the military, and some college food service operations is to create a menu that fulfills specific dietary requirements. In a simplified version of this problem, suppose that a dietitian is trying to create a meal that provides a person’s total daily requirement of vitamin C and iron by serving exactly two dishes—potatoes and spinach. For a female between the ages of 19 and 24, the recommended daily allowance of vitamin C is 60 milligrams (mg) and that of iron is 15 mg. The dietician needs to determine how many potatoes and how much spinach a 21-year-old female must eat per day to fulfill these requirements. {INSERT Photo 3.1} When two linear equations must be solved at the same time, they constitute a system of simultaneous equations. To solve the system of equations, we must find values of both X and Y that simultaneously satisfy both equations. There are two methods for solving such problems, one graphical and the other algebraic. We begin with an example of the graphical approach. A GRAPHICAL SOLUTION: — INTERSECTION OF TWO LINES To solve a system of equations, we plot each equation on the same coordinate grid and estimate the point of intersection. This employs the key insight needed to solve two equations simultaneously: in order for a point to satisfy both equations, it must lie on both lines. Only a point at the intersection of the two lines can satisfy both equations simultaneously. Example 1 Solving a System of Two Equations Graphically Solve the following two equations simultaneously by the graphical method: A. 2X + 5Y = 20, B. 4X – 3Y = 14 Solution Linear equations in the general form are easy to graph because with a simple calculation one can find the vertical and horizontal intercepts. Starting with equation A, 2X + 5Y = 20, we find the vertical intercept by setting X = 0 and substituting that into the equation: 2(0) + 5Y = 20, 5Y = 20, Y = 4. So the vertical intercept is the point with coordinates (0, 4). Similarly, we find the horizontal intercept for equation A by setting Y = 0 and solving, 2X + 5(0) = 20, 2X = 20, X = 10. So the horizontal intercept is (10, 0). We graph the line 2X + 5Y = 20 by plotting the points (0, 4) and (10, 0) and connecting the points with a straight line. By repeating the steps above for equation B, we determine its vertical and horizontal intercepts. Using the equation 4X – 3Y = 14, we find the intercepts to be (0, <->14/3) and (7/2, Chapter 3 Systems of Linear Equations and Inequalities 3 0), respectively. Again, we graph the line by plotting these points and connecting them with a straight line. The graph of these two equations appears in Figure 3.1.1. The solution to the system of equations is the point where the two lines intersect. By inspection, it appears to have an Xcoordinate of about 5 and a Y-coordinate of about 2. Thus, the solution to the system of equations should be close to X = 5, Y = 2. {INSERT FIG. 3.1} Learning with Technology Finding the Intersection of Two Lines One of the special features of a graphing calculator is its capability of finding the intersection of two lines or curves. Although the graphing calculator is optional with this textbook, a word here about its use will help some students. A graphing calculator can plot several equations at once, but each must be given to the calculator in the functional form “Y = . . . .” To obtain a graphical solution to the system of equations above using a graphing calculator, we must rewrite the equations in this fashion. The first equation in Example 1 was 2X + 5Y = 20. We subtract 2X from both sides and divide through by 5 to get Y = 4 – 0.4X. Doing the same with the second equation, taking care when dividing through by <->3, the system becomes A. Y = 4 – 0.4X, B. Y = <->14/3 + 4/3X. Now you are ready to use the graphing calculator, following these steps: 1. Enter these two equations as Y1 and Y2 or on any two other available lines in the Y = menu. Clear any other equations that might have been selected for plotting the last time you used the calculator and select these two alone. {INSERT FIG. UN3.1a} 2. Find a graphing window that will display these lines and their intersection. The slope-intercept forms in A and B reveal the vertical intercepts of each graph. A vertical screen interval of [<->5,4] will show both intercepts. What horizontal screen interval will display a complete graph? Setting Y = 0 in each equation we do a quick calculation, as in Example 1, to find the horizontal intercepts: X = 10 for the first equation and X = <->7/2 for the second. The horizontal viewing window should therefore extend as far as 10 to the right. The vertical intercepts correspond to X = 0, so all intercepts will appear if we use a horizontal window [0,10]. {INSERT FIG. UN3.1b} 3. Make sure that both equations are selected for graphing, and then call up the graph. It is the same as Figure 3.1, but it should appear on the calculator as shown here. Most calculators do not show the numbers on the axes, but you can infer them from the tick marks and the window boundaries. {INSERT FIG. UN3.1c} 4. Most graphing calculators can find the intersection of two lines. A CALC INTERSECT routine in the Texas Instruments models will find the intersection. {INSERT FIG. UN3. 1d} Chapter 3 Systems of Linear Equations and Inequalities 4 If your calculator does not have an INTERSECT routine, the TRACE feature will still allow you to move along one line until you get to the intersection with the other line. Toggle back and forth between the two lines until you find the point where the Y-values for each equation are as close as possible. ZOOM in on the point of intersection and TRACE to get on one of the lines and find a more accurate estimate. These show that the solution to the system of equations is X = 5 and Y=2. {INSERT FIG. UN3.1e} AN ALGEBRAIC SOLUTION: THE SUBSTITUTION METHOD An algebraic way to solve a system of linear equations is called the substitution method. This will give a more precise solution than the graphical method above. We’ll use this method to solve the problem stated in Example 1 above. Example 2 Solving a System of Two Equations Algebraically Solve the system from Example 1 algebraically: A. 2X + 5Y = 20, B. 4X – 3Y = 14. Solution 1. Pick one of the two equations and solve for one of the variables. The two equations in Example 1 are both written in the general form, so it doesn’t matter which one we pick or which variable we solve for. Here we’ll begin by solving for X in equation A. Starting with 2X + 5Y = 20, we subtract 5Y from both sides, 2X = 20 – 5Y, and divide through by 2, X = 10 – 5 Y. 2 2. Use this formula for X in the other equation. We substitute the formula 10 – place of X in equation B and then solve for Y: 4X – 3Y = 14, 4(10 – 5 Y) – 3Y = 14, 2 40 – 10Y – 3Y = 14, <->13Y = <->26, Y = 2. This gives us one part of the solution, Y = 2. 5 Y in 2 Chapter 3 Systems of Linear Equations and Inequalities 5 3. Substitute this known value for Y into either equation A or B, and calculate the corresponding value of X. Equation A was already put into slope intercept form for X. Using that we get X = 10 – 5 Y, 2 X = 10 – 5 (2), 2 X = 5. The answer X = 5 and Y = 2 was suggested as an approximate solution in our graphical analysis of Example 1, which shows that the graphical solution can be a good double check of our calculations in the algebraic method. To verify the solution algebraically, we substitute the values for X and Y into equation B: 4X – 3Y = 14, 4(5) – 3(2) = 14 . Make a Note Substitution Method in Three Steps 1. Pick one of the two equations and solve for one of the variables. 2. Substitute the formula for that variable into the other equation wherever that variable appears and solve for the remaining variable. 3. Substitute the value you obtained in Step 2 into the equation you got in Step 1 and calculate the value of the first variable. Example 3 Designing a Diet Let’s return to the application introduced in the Spotlight at the beginning of this chapter. The problem was to design a diet consisting of only two foods, spinach and potatoes. The objective of the diet is to satisfy two minimum daily requirements: 60 mg of vitamin C and 15 mg of iron. To plan the diet, we need to know how much of each nutrient is found in a pound of each type of food. The table shows the nutritional values of each. * Nutrient Vitamin C (mg) Iron (mg) Per pound of . spinach 45 16 Per pound potatoes 58 1.5 The quantities of spinach (S, lbs.) and of potatoes (P, lbs.) that fulfill the vitamin C requirement should satisfy an equation that reads as follows: The amount of vitamin C from spinach and the amount from potatoes must sum to 60 mg. Each pound of spinach gives 45 mg of vitamin C, and each pound of potatoes gives 58 mg, so the statement above translates into mathematical form as *Nutritive Value of Foods, Home and Garden Bulletin, No. 72 (Washington, DC: US Government Printing Office). < http://www.nal.usda.gov/fnic/foodcomp/Data/HG72/hg72.html> Chapter 3 Systems of Linear Equations and Inequalities 6 A. 45S + 58P = 60. By the same logic, we may also write an equation that expresses the iron requirement of 15 mg using the data on iron content from the table, B. 16S + 1.5P = 15. How can we find a value for S and a value for P that satisfy both of these equations at the same time? Solution Graphical Solution To get started on such problems, it always helps to draw a picture. Figure 3.2 shows a graph of both equations on the same chart. We do so by plotting the vertical and horizontal intercepts of each equation and connecting the points by a straight line as we did in Example 1. The intercepts (S, P) of equation A are (0, 60/58) and (60/45, 0). Those of equation B are (0, 10) and (15/16, 0). {INSERT FIG 3.2} Think about This Examine Figure 3.2, and see if you can answer this question yourself: Are there any combinations of S and P that can satisfy both equations at the same time? Remember that only at the point of intersection are both equations satisfied simultaneously. In Figure 3.2, it looks as though the coordinates of the point of intersection are approximately S = 0.9 pounds of spinach and P = 0.5 pounds of potatoes. But an eyeball guess about the coordinates of the intersection point won’t be precise enough for a dietician, who is responsible for meeting dietary requirements. So let’s see how to get an exact answer. Algebraic Solution The first step is to isolate one of the variables. Neither the vitamin C nor the iron equation is in slope-intercept form, so it does not matter which variable we isolate or in which equation. We shall isolate the variable S in the vitamin C requirement, equation A, A. 45S + 58P = 60. Subtract 58P from both sides, 45S = 60 – 58P. Divide through by 45, S= 60 58 – P, 45 45 S = 1.333 – 1.289P. Make a Note Using Decimals and Rounding Answers In the real world, the solutions to math problems rarely come out to be integers or simple fractions. They are real numbers, which we approximate as decimals. It is the common practice of applied mathematicians to round their answers to the same number of significant digits as the data given in a problem, except where an answer calls for dollars and cents. We will follow that practice here. However, during the calculations toward a solution, we will carry at least one extra digit. Otherwise, several small rounding errors might add up to a larger error in the final solution. The next step is to substitute this formula wherever S appears in the iron requirement, equation B, and then solve for the remaining variable, P. We start with the iron equation as it was originally stated, B. 16S + 1.5P = 15. Now, substitute the formula 1.333 – 1.289P for the variable S in this equation, 16(1.333 – 1.289P) + 1.5P = 15. Chapter 3 Systems of Linear Equations and Inequalities 7 Think about This Be sure you understand how this equation was created. It is a combination of equations A and B above, formed by putting the formula for S from equation A into the place where S was in equation B. Expand the parentheses using the distributive rule, 21.333 – 20.624P + 1.5P = 15, and combine the terms in P, <->19.124P = <->6.333. Divide through by <->19.124, and retain two significant digits in the answer (because the data in the problem were stated in two significant digits), P ≈ 0.33 pounds of potatoes. This is the value of one variable. Now we must find a value for the other variable. Think about This Knowing P now, how can we find the corresponding value of S? For this purpose, we go back to the first equation. We substitute this value for P into the version of equation A in which the variable S was isolated: S = 1.333 – 1.289P, S = 1.333 – 1.289(0.33), S = 1.333 – 0.425, S ≈ 0.91 pounds of spinach (to two decimal places). The solution to the problem is that a daily diet of 0.91 pounds of spinach and 0.33 pounds of potatoes will provide the required amounts of both vitamin C and iron. There are two ways to check the correctness of this solution, one graphical and one algebraic. The graphical way to check the answer is to sketch a plot of the two equations and see if the intersection of the lines appears to be near P = 0.33 and S = 0.91. For this we may refer back to our graphical solution in Figure 3.2. The lines do appear to intersect near S = 0.9 and P = 0.3. The algebraic way to check this solution is to substitute these values of S and P into each of the two original equations. We want to see if the calculation results in the appropriate righthand-side value of the equation—at least to within a tolerable error (third significant digit) due to rounding off our divisions in the process of calculation. Start with the vitamin C equation: A. 45(0.91) + 58(0.33) = 60? 40.95 + 19.14 = 60? 59.99 ≈ 60; close. Now check the iron requirement: (b) 16(0.91) + 1.5(0.33) = 15? 14.56 + 0.495 = 15? 15.055 ≈ 15; close. We conclude that to get the daily required amounts of vitamin C and iron on a diet of potatoes and spinach, a woman aged 19-24 would need to eat about 0.91 pounds of spinach and 0.33 pounds of potatoes per day. There are numerous important applications of systems of linear equations in business and economics. Let’s take a look at some of these. Chapter 3 Systems of Linear Equations and Inequalities 8 Example 4 Making Production Decisions in a Wood Shop Let’s say that a wood shop produces only two products, bookcases and chairs. These two outputs require the same two inputs, labor time and machine time. This table gives the necessary information about how much of each input is used to create one unit of each product, for example, that the manufacture of one bookcase requires 0.75 labor hours and 0.5 machine hours. Input requirements Per bookcase Per chair Labor 0.75 2.5 Machine 0.5 1.0 The owner of the workshop has 2 employees who each work a 40-hour week. In addition, his machinist runs the machine for 36 hours a week and does maintenance on the machine for the remainder of the 40-hour workweek. How many bookcases and chairs can the shop make if it uses all the labor and machine time available? Solution FORMULATING THE PROBLEM To set up this problem mathematically, we write two equations in two unknowns. (See Fig. 3.3). The first equation will say in words, “The number of hours needed to make B bookcases plus the hours needed for C chairs must equal 80 hours.” The second is analogous for machine hours. A. 0.75B + 2.5C = 80 (labor hours) B. 0.5B + C = 36 (machine hours) {INSERT FIG. 3.3}. ALGEBRAIC SOLUTION 1. Isolate one of the variables in one of the equations. In the machine equation B, the variable C already has a coefficient of 1, so that makes our task a little easier: B. 0.5B + C = 36 (machine hours). Subtract 0.5B from each side: C = 36 – 0.5B. 2. Substitute this expression for C into the labor equation A and solve for the remaining variable: A. 0.75B + 2.5C = 80 (labor hours), 0.75B + 2.5(36 – 0.5B) = 80, 0.75B – 1.25B + 90 = 80, <->0.5B = <->10, B = 20 bookcases. 3. Substitute this value for B into the isolated version of equation B: C = 36 – 0.5(20), C = 36 – 10, C = 26 chairs. Chapter 3 Systems of Linear Equations and Inequalities 9 In conclusion, when using all the labor hours and machine time available, namely 80 labor hours and 36 machine hours, the shop can make 20 bookcases and 26 chairs. We check these solutions by substituting them into the original equations: A. 0.75B + 2.5C = 80, 0.75(20) + 2.5(26) = 80? 15 + 65 = 80? 80 = 80 exactly. B. 0.5B + C = 36, 0.5(20) + (26) = 36? 10 + 26 = 36? 36 = 36 exactly. Learning with Technology Find the size of the window needed to view the graphs of equations A and B. Put equations A and B in slope intercept form as “C = . . . ,” but keep the constants in fractional form. Hint: Equation A would be written as C = 80/2.5 – 0.75/2.5*B. Graph the equations and find their intersection using the INTERSECT or TRACE routine. How would the solution change if there were 40 machine hours available rather than 36? What if there were 50? What if 60? Application Market Supply and Demand The “market” for any product consists of buyers and sellers. The buyers’ need for the product is described by a demand curve (or demand equation). This is a graph that shows the relationship between the price (P) in the market and the quantity (Q) of the good that would be demanded (purchased) by buyers. It is a general economic principle that when the price is higher, people consume less of a product. Graphically, this means that the demand curve slopes downward. The supply side of a market is described by a supply curve (or supply equation) that indicates what quantity (Q) would be produced by suppliers and offered in the market at each level of price (P). The supply curve slopes upward because producers require a higher price if they are to increase production output within the capacity of their existing factories. The reason is that when a factory has to produce a lot more units of a product in the same amount of time, crowding, mistakes, and inefficiencies tend to result in a higher cost of production. At a price where the quantity demanded is equal to the quantity supplied, the market is said to have “cleared.” In other words, the combination of price and quantity in the market has to fit both the buyers’ demand equation and the producers’ supply equation. The combination of price and quantity that makes this possible is called the market equilibrium. In other words, the P and Q in the market equilibrium problem have to satisfy two different equations at the same time. Example 5 Demand for CD Players Many electronics companies sell CD players on the Internet. Suppose that in their experience, customers will not pay more than $200 for a particular model of CD player. But customers would be willing to buy a player (in any 1 year) if the price is lowered. Let’s say that to sell Chapter 3 Systems of Linear Equations and Inequalities 10 one more player per year, no matter how many players are currently selling, the price must come down by $0.006. Create the demand equation that describes the relationship between P and Q and plot it. Solution According to this description, there will be a linear relationship between quantity sold in the entire market and the price of a CD player. The demand line will have a vertical intercept of $200 (where Q = 0) and a slope of <->0.006. The equation representing this demand curve would appear in slope-intercept form as P = 200 – 0.006Q. Using this market demand equation, you can calculate the corresponding value of P if you know the number Q of players that need to be sold in a year. Likewise, given a known value for P, you can solve for the corresponding value of Q to find out how many players will be sold at that price. Practice 1. If 10,000 players need to be sold in a year, what price should the retailers set? 2. How many players would customers want to buy if the price is $100? Answers: (1) Use the demand equation P = 200 – 0.006Q with Q = 10,000, getting P = $140. (2) Solve the demand equation P = 200 – 0.006Q with P = 100, and find Q ≈ 16,667 players. The market demand equation describes a negative association between P and Q. The higher the Q, the lower the P. This is seen in the downward-sloping graph of Figure 3.4, and it is evident from the negative slope coefficient in the equation P = 200 – 0.006Q. {INSERT FIG 3.4} The supply equation tells a very different story about P and Q. In many industries, the production cost per unit tends to increase as the production level increases, due to production inefficiencies, overtime pay to workers, and so on. Thus, manufacturers will supply more of a product only if the price is higher. In other words, the higher the Q, the higher must be the P. This is characteristic of a positive association between the variables P and Q. We should expect that the supply equation will have a positive slope coefficient and that its graph will have an upward slope. Another feature of the production process is that, from the manufacturer’s point of view, there is a minimum price below which producing new players is not profitable, so no players will be produced. In graphical terms, this means that the supply line will have a vertical intercept at some positive value of P. Example 6 Supply of CD Players Suppose that in our CD player example it costs the manufacturers of these players a minimum of $50 to produce each player. For each additional batch of 1,000 players that they have to produce in a given year, the production cost per player goes up by $4, meaning $0.004 per player. The production cost is the minimum price at which the manufacturer would be willing to supply the players. Write an equation that describes the manufacturer’s willingness to supply CD players and plot it. Chapter 3 Systems of Linear Equations and Inequalities 11 Solution Again, we have a linear relationship between P and Q. The supply line will have a vertical intercept at $50, where Q = 0, and a slope of 0.004. Thus, the equation describing the manufacturers’ price in relation to the number of players produced can be written as P = 50 + 0.004Q. Just as with the demand equation, if you are given a value for Q, you can calculate the corresponding P very quickly. Or, given a value for P, you can solve for the corresponding value of Q rather easily. But the supply equation describes a very different relationship between P and Q. As we see in Figure 3.5, it has an upward slope. Recall that the demand equation in Figure 3.4 sloped downward. {INSERT 3.5} In the Practice box, we calculated that customers would purchase 16,666 units at a price of $100. Now let’s find out how many players the manufacturers would be willing to produce if the price is $100. We solve the supply equation, P = 50 + 0.004Q with the value of P set at $100, 100 = 50 + 0.004Q, 50 = 0.004Q, Q = 12,500 players. These two calculations show that at a price of $100 per player, there is a demand for 16,666 units, but the manufacturers are willing to produce only 12,500 units. When a market has an excess of demand over supply (or when supply exceeds demand) it is said to be in disequilibrium. What combination of price and quantity will balance supply and demand? It will be a level of P and a level of Q that satisfies both the demand equation and the supply equation simultaneously. Graphically, the market equilibrium will be a point with coordinates (P, Q) that lies simultaneously on the market demand line and on the supply line. That is the point of intersection of the two lines. Example 7 Market Equilibrium of CD Players Determine the point at which demand meets supply in our CD player application. Solution GRAPHICAL SOLUTION By plotting both the demand and supply equations from the previous examples on the same graph, we can see that the equilibrium will correspond to Q ≈ 15,000 players at a price of P ≈ $120 each in Figure 3.6. {INSERT FIG. 3.6} ALGEBRAIC SOLUTION Algebraically, the market equilibrium point (P, Q) is the pair of numbers that satisfy simultaneously the demand and supply equations, A. P = 200 – 0.006Q (demand equation), B. P = 50 + 0.004Q (supply equation). Chapter 3 Systems of Linear Equations and Inequalities 12 To find the equilibrium algebraically, we solve this system of two equations simultaneously, and for that we use the substitution method. Each of these equations is already in slope-intercept form, so they have one variable isolated on the left side of the equality sign. It does not matter which equation we use for the substitution. Let’s take the demand equation. We substitute the formula 200 – 0.006Q everywhere the letter P occurs in the supply equation, P = 50 + 0.004Q (supply equation). This gives us a new equation, (200 – 0.006Q) = 50 + 0.004Q. We now solve this for the unknown value of Q. Add 0.006Q to both sides: 200 = 50 + 0.010Q. Subtract 50 from both sides: 150 = 0.010Q. Divide through by 0.010 to get Q: Q = 15,000. To find the corresponding value of P, we may use either equation A or B. Substitute the value Q = 15,000 into it, and solve for P. We take equation A: P = 200 – 0.006(15,000), P = 200 – 90, P = $110. In summary, the solution to the system of equations is the pair of values Q = 15,000 and P = $110 as Figure 3.6 confirms. In other words, the point where the two lines intersect in the QP plane has coordinates (15,000, 110). This point corresponds to the production and consumption of 15,000 CD players at a market price of $110 each. PARAMETER ESTIMATION To write a linear equation, we have been using large letters for the variables and small letters for the constants or parameters, which will be specific numbers when the equation is actually used. The slope-intercept form of the linear equation, Y = mX + b, has two parameters: the coefficient of X (the slope parameter, m) and the constant term, b. In most of the application problems we encounter, the linear equation describes a general rule that shows the relationship between the variables. Giving known values to the parameters in the equation makes that relationship specific, as when we define a particular demand equation to be Y = 200 – 0.006X. As you have seen, the task in most problems is to find a value for an unknown variable when we already know the parameters of the equation. However, one of the most important applications of systems of equations is to find the parameters that define a specific relationship between two variables. To find the two parameters for a linear equation, we have to use information about the relationship between the two variables in two or more observed cases. In the applications we will examine below, we will use exactly two cases. Two cases give us two points of data, and two points determine a line. This is enough to give us a specific estimate of the two parameters in a linear equation. Chapter 3 Application Systems of Linear Equations and Inequalities 13 A Formula for the Price of a House Two of the most important factors in determining the price of a house are its square footage of floor space and the amount of land that comes with it. In other words, the price P of a house is a function of its square footage S and the amount of its land, L. In general, the price of a house will go up linearly with its square footage and likewise with the area of land on which the house sits. This means that the general form of an equation for the price will be house price P = cS + dL, for some parameters c and d. The constants, or parameters as they are often called, indicate respectively the number c of dollars paid per square foot and the number d of dollars paid per acre. When people buy a house and land, it is typically bought as one package and paid for with one sum of money. But the tax laws treat the building and the land very differently. If you buy a house in order to rent it out, you can depreciate part of the value of the building every year (that is, treat it as a business expense), but you cannot depreciate the value of the land. The logic of the Internal Revenue Service is that buildings wear down over time, so it is reasonable to allow businesspeople a tax deduction for that, but land does not wear down. Therefore, anyone who buys a house to rent it out wonders, “How much of the purchase price is attributable to the house and how much to the land?” This question can be answered by setting up a system of two equations, each of which has two unknowns. This problem has an interesting twist, like the problem of finding a commission rate for a salesperson. The unknowns in the system of equations are going to be the constants c and d in the house price formula, and the constants are going to be the actual characteristics of two houses that have sold for known prices. If you can mentally tolerate this turn of the tables for another minute, you’ll see how useful this idea is. Example 8 Price of Land and Buildings Suppose that we have real data on two houses that recently sold in town. The data are summarized here in the table. Property Size of the building Land Market price (sq. ft.) (acres) 110 Main St. 900 0.25 $105,000 412 St. 1,200 0.40 $145,000 Dexter As surprising as it may seem, this is all the information we need in order to determine how the price of a property is calculated from the square footage of the house and the acreage of the land. Using this table, write two equations in the two unknown parameters. Solution It is often helpful to first look for an equation that can be stated in words. In this example, we will set up two equations that each state, The cost of the building plus the cost of the land equals the cost of the property. Assuming that in the town, the general price of the building is c dollars per square foot, then the cost of the building on property A must be c 900. Likewise, if the cost of land is d dollars per acre, then the 0.25 acres at 110 Main Street must have cost d 0.25. The sum of these must equal the price paid for that property. This logic gives us the first of the following equations. The second equation is obtained similarly: Chapter 3 Systems of Linear Equations and Inequalities 14 A. c900 + d 0.25 = 105,000 (price of 110 Main St.), B. c1,200 + d 0.40 = 145,000 (price of 412 Dexter St.). Notice what we have done here. In each equation, we used actual data for the “variables” S and L but left the parameters c and d unspecified. This gives us a linear equation in which the unknowns are the parameters c and d. This is an example of how real data are used to estimate the parameters of a mathematical model (an equation) for a phenomenon like house prices. We can solve these two linear equations using the substitution method. Let’s take the first equation and isolate the unknown d. Starting with c 900 + d 0.25 = 105,000, we get d = 420,000 – c 3600. Substitute this into the second equation, and solve for the unknown parameter c: c1200 + (420,000 – c 3600)0.40 = 145,000. Expand the parentheses, 1,200c – 1,440c + 168,000 = 145,000. Combine like terms, <->240 c = <->23,000, and divide through, c ≈ $95.83 per square foot. Our equation for the parameter d was d = 420,000 – c =3600. Substitute the known value for c back into our equation for the parameter d and solve. d ≈ 420,000 – (95.83)3,600. d ≈ $75,012 per acre. There we have it. The actual data for two properties have enabled us to determine the two unknown parameters of a linear equation that describes the price of a property as a function of size of the building and the amount of the land. To check these answers graphically, we plot the two original equations and see if the intersection corresponds with the solutions we just found. Figure 3.7 shows that we should expect a solution that has a price per acre somewhat below $100,000 and a price per square foot a little below $100. The solution c = $95.83 per square foot and d = $75,012 per acre is consistent with this picture. {INSERT FIG. 3.7} To check the answers algebraically, we substitute our values of c and d into the original equations and see if the equations hold up. The equation for 110 Main Street is A. c 900 + d 0.25 = 105,000 (price of 110 Main St.). Substitute the solutions c = 95.83 and d = 75,012 into this equation and see if the left and right sides match up: (95.83) 900 + (75,012) 0.25 = 105,000, 86,247 + 18,753 = 105,000, Chapter 3 Systems of Linear Equations and Inequalities 15 105,000 = 105,000. [ED: code?] Now do the same for 412 Dexter Street, B. c 1,200 + d0.40 = 145,000, (95.83) 1200 + (75,012)0.40 = 145,000, 114,996 + 30,004.80 = 145,000. 145,000.80 ≈ 145,000 (close enough). Our purpose in searching for Please restore all in this sentence}the costs per square foot and per acre was to write an equation for the cost of a property if the size of the building and the amount of the land were unknown. We have found those cost parameters, so we write them into the template for our linear equation for the property price, P = $95.83S + $75,012L, where S is the size of the house, and L is the amount of land. Using this equation, we can now determine the separate value of the building and of the land in each of the houses for which we have data. Consider the property at 110 Main St. For tax purposes, the value of the building would be $95.83(900) = $86,247, and the value of its land would be $75,012(0.25) = $18,753. It is quite remarkable that a few points of data enable us to tease out information that seemed at first to be buried inextricably in the situation. You will find that to be true in the examples that follow. Application Estimating Productivity Coefficients A production manager in a firm needs to know how long it takes employees to produce different products. One way would be to observe employees in the process of manufacturing each product and to measure the time they take on each type, but employees do not usually like to be watched. There is another way, though, which does not require someone to observe people working on each type of product individually. A manager can simply count the number of units of each type of product produced in, say, an 8-hour day. If there are only two products and the manager has this information on two days, it will be possible to calculate the time it takes to produce one unit of each type of product. How can a manager do this? Let’s assume that there are two products, labeled 1 and 2 and that there is an unknown amount of time required to produce each product. We write c1 for the number of hours needed to manufacture a unit of product 1 and c2 for the number of hours needed to manufacture a unit of product 2. The total time that will be spent in manufacturing a quantity Q1 of product 1 will be the number of hours required to manufacture a single unit multiplied by the number of units produced. That total is c1Q1. Likewise, the total time that will be spent in manufacturing a quantity Q2 of product 2 will be c2Q2. The mathematical model for the total number of hours (H) a worker needs to manufacture Q1 units of product 1 and Q2 units of product 2 is therefore expressed by the equation, H = c1Q1 + c2Q2, otherwise the variable H is defined but note used? OK} Example 9 Employee Productivity Let’s introduce some values to make the application concrete. Suppose we know the following: Chapter 3 Systems of Linear Equations and Inequalities 16 Day 1: During this 8-hour day an employee made 6 units of product 1 and 8 units of product 2. Day 2: During this 8-hour day, the same employee made 3 units of product 1 and 12 units of product 2. We will write one equation using each day’s information. In plain English, the equations will state: The number of hours used to manufacture product 1 plus the number of hours used to manufacture product 2 is equal to the 8 hours in a worker’s day. In mathematical form, this reads as A. c1 6 + c2 8 = 8 hours worked in the first day, B. c1 3 + c2 12 = 8 hours worked in the second day. Find out how much time it takes to make 1 unit of each product by solving this system of simultaneous linear equations. Solution GRAPHICAL SOLUTION These two equations are in general form, so it should be easy to do so in paper and pencil by finding the vertical and horizontal intercepts of each line and then drawing the lines. The intercepts of line A are at (0, 1) and (4/3, 0). The intercepts of line B are at (0, 2/3) and (8/3, 0). A plot of these two lines on a window [0,3] by [0,1] appears in Figure 3.8. {INSERT FIG 3.8} The dashed lines show that the solution to the system of equations ought to be production times of approximately c1 = 2/3 of an hour per unit of product 1 and c2 = 1/2 hour per unit of product 2. ALGEBRAIC SOLUTION Since we are accustomed to solving systems in which the unknown appears to the right of the specified coefficient, we may use the commutative law for real numbers to rearrange the order of multiplication. This makes the system of equations appear as follows: A. 6c1 + 8c2 = 8 , B. 3c1 + 12c2 = 8. The first step is to isolate one of the unknowns in one of the equations. We choose to isolate c2 in equation A because when we divide through by the coefficient of c2, which is 8, we will get familiar fractions and numbers: A. 6c1 + 8c2 = 8, 8c2 = 8 – 6c1, c2 = 8 6 – c1, 8 8 c2 = 1 – 3 c1. 4 The second step is to substitute that expression for c2 into equation B. 3c1 + 12c2 = 8, 3c1 + 12(1 – 3 c1) = 8. 4 Chapter 3 Systems of Linear Equations and Inequalities 17 With a little arithmetic and the rules of algebra, we solve this for the unknown c1. 3c1 + 12 – 9 c1 = 8, <->6c1 = <->4, c1 = 2 . 3 In the final step, we substitute this value for c1 back into our equation for c2 to find the remaining unknown, c2 = 1 – =1– = 3 2 4 3 2 4 1 . 2 Our solution to these equations reveals that it takes 2/3 of an hour (or 40 minutes) to produce each unit of product 1 and 1/2 of an hour (or 30 minutes) to produce each unit of product 2. These are exact solutions, and they correspond well to the eyeball estimates taken from the graphs in Figure 3.8. Application Estimating the Effects of Advertising Marketing managers often complain, “Half my advertising budget is wasted. I know that. The problem is, I don’t know which half.” A marketing manager needs to find out exactly what are the effects of an advertising campaign. However, it is usually not practical to take out one advertisement at a time in order to see its effects clearly. Typically a manager tries many advertising tactics at once. But then it can be a real problem to sort out the individual effects of all the separate advertisements and other marketing tactics. For that purpose, it helps enormously to know how to solve systems of equations. With data about the sales of a product over 2 months, as a result of advertisements of 2 different types, you can usually determine how much of an effect each of the advertisements was having. Example 10 Terry’s Water Purifiers While still in college, Terry Reilly was given the Midwest dealership for a brand of water purifiers, which he sells from his dorm room as a way to make extra money. His only advertising has been a series of 30-second radio ads. He runs them on the campus radio station, WABC, and on a station in town, WXYZ. The number of ads he ran per week and the results in each of two weeks are summarized in the table. Week Ads on Ads on Units WABC WXY Sold Z 1 15 20 55 2 10 15 40 Determine the effectiveness of running ads at each station. Solution Chapter 3 Systems of Linear Equations and Inequalities 18 With two radio stations and two weeks of data, it should be possible to tease out the separate effects of the ads. We will assume that for each station there is a proportional relationship between number of ads run and the number of units sold per week. To write this idea in an algebraic form, let the letter A represent the number of ads run per week in WABC and let X be the number of ads run per week at WXYZ. Let c be the number of units sold per advertisement on WABC and write d for the number of units sold per advertisement on WXYZ. The verbal model, in short, says, The sales from ads in WABC and the sales from WXYZ make up the total number of purifiers sold. This can be written algebraically as cA + dX = total purifiers sold. We use each of the two weeks’ data for A, X, and the number of purifiers sold in this algebraic model to give us two equations: c15 + d20 = 55 (units sold, week 10 c10 + d15 = 40 (units sold, week 2). This is a system of two simultaneous linear equations in two unknowns, c and d. GRAPHICAL SOLUTION To graph these two equations on the same coordinate system, we should estimate the intercepts of each equation and choose our window appropriately. In rough numbers, for the first equation, when c = 0, d ≈ 3, and when d = 0, c ≈ 3. Likewise, for the second equation, when c = 0, d ≈ 3, and when d = 0, c = 4. So an interval of [0,4] should work for c and [0,3] should work for d. Figure 3.9 shows the graphs of these equations on that window. The lines corresponding to the two equations have nearly the same slopes and intercepts, so it is a bit difficult to see exactly where they intersect. It appears to be near c = 1 and d = 2. {INSERT FIG. 3.9} ALGEBRAIC SOLUTION How should we start solving this system? For step 1, let’s isolate the c in the second equation because its coefficient, 10, might help keep the numbers simple for a while when we have to divide through. Subtracting d 15 from both sides, we get c 10 = 40 – 15 d. Dividing through by 10 yields an equivalent form of the second equation, c = 4 – 1.5 d. As step 2, we substitute this expression for c into the other equation, and solve for d: (4 – 1.5d)15 + d 20 = 55, 60 – 22.5d + 20d = 55, 5 = 2.5d, d = 2 sales per ad at WXYZ. As step 3, we substitute that value of d back into the equation for c, c = 4 – 1.5 2, Chapter 3 Systems of Linear Equations and Inequalities 19 c = 1 sale per ad at WABC. As a check, we put these values for c and d back into the original equations to see if they produce equality. They are not used for double checks above. If you feel they are necessary, OK to use "Question:" instead? (1) 15 + (2) 20 = 55 (units sold, week 1), 15 + 40 = 55;✓ (1) 10 + (2) 15 = 40 (units sold, week 2), 10 + 30 = 40. ✓ This analysis shows that each ad at WABC results in the sale of 1 water purifier, and each ad at WXYZ results in 2 sales. PRACTICAL IMPLICATIONS OF THE MODEL Does this solution mean that Terry should advertise only in WXYZ, where he gets 2 sales per advertisement? Not necessarily. Terry should want to advertise where he gets the greatest return on his money, which means the highest sales per dollar of advertising. Therefore, he has to consider both the effectiveness of the ad (units sold per ad) and the cost of each ad. Suppose the spots on WABC cost $10 each and the spots on WXYZ cost $25. Which radio station had the higher ratio of sales to advertising cost? Terry knows from the previous analysis that each ad at WABC produces 1 sale. This means that he sells 1/$10 = 0.10 purifiers per dollar of advertising spent there. At WXYZ, each ad produces 2 purifier sales, so he sells 2/$25 = 0.08 purifiers per dollar spent there. In this example, WABC produces the higher number of sales per dollar, even though the number of responses per ad is lower. Make a Note Solving a Linear System for Variables or for Parameters The template for the general form of a linear equation is aX + bY = c. This can be made into a specific linear equation in either of two ways. 1. When specific parameters are known, the equation defines a relationship between the variables X and Y (e.g., the wood shop problem). A system of two such equations can be solved to find values for X and Y. 2. When the parameters are unknown, specific values for X and Y and for the right-hand side number c turn the template into a linear equation with unknown parameters. Data for two combinations of X, Y, and c will lead to two linear equations that can be solved to find the parameters a and b. TWO UNUSUAL CASES In the examples that we have seen so far in this section, the two equations corresponded to two distinct lines that intersected in a single point. The coordinates of that point of intersection gave the values of the two variables that solved the system of equations as a whole. This principle is quite general. The solution to the system of equations is always the set of points in the plane that lie on both lines. But does every pair of lines intersect in a single point? Think about This Is it possible to draw one or more examples of pairs of lines that do not intersect in a single point? Chapter 3 Systems of Linear Equations and Inequalities 20 There are two ways in which the ordinary situation of a single-point solution can fail. We call these cases “unusual” because they differ from what one would expect in a system of two linear equations in two unknowns. An Inconsistent System A system of equations is called inconsistent if it has no solution. Two linear equations will be inconsistent when they do not have a point of intersection. In this case, they must correspond to two parallel lines (Fig. 3.10). {INSERT FIG. 3.10} Algebraically, how can we recognize that a system of two linear equations corresponds to two parallel lines? The key idea is that parallel lines have the same slope. So, when two linear equations are written in the slope-intercept form, it will be obvious whether or not their slopes are the same. To be parallel, the lines must also have different vertical intercepts. The slope-intercept form will also reveal whether or not the intercept parameters are different. The graph shown in Figure 3.10 corresponds to a system of equations that, when written in slope-intercept form, appears as, Y = 5X + 7, Y = 5X + 3. The slope parameter is the coefficient of the X variable, which is 5 in both equations. Therefore, these equations represent lines that have the same slope. The vertical intercept is +7 in the top equation and +3 in the bottom equation, showing that the lines are in fact distinct from each other. Not all systems of linear equations come to us in slope-intercept form, so it is not always possible to recognize an inconsistent system of equations by inspection. However, if you try to work the substitution method on an inconsistent system, you get an interesting result. What happens? Let’s see by working an example. Example 11 An Inconsistent System Consider the following system of two linear equations in two unknowns, A. 2Y = X – 4 B. Y + 7 = 0.5X. Try to solve this system using the steps in the substitution method. Solution 1. Isolate one variable in one of the equations. The bottom equation is closest to slopeintercept form, so let’s isolate the Y by subtracting 7 from both sides. This gives us the equivalent system of equations. A. 2Y = X – 4, B. Y = 0.5X – 7. 2. Substitute the expression for that variable into the other equation. We put (0.5X – 7) in place of Y in the top equation, and simplify to solve for X. The top equation becomes A. 2(0.5X – 7) = X – 4. Chapter 3 Systems of Linear Equations and Inequalities 21 Expand the parentheses, 2 0.5X – 2 7 = X – 4, and do the arithmetic, X – 14 = X – 4. Add 4 to both sides, X – 10 = X. Simplify by subtracting X, <->10 = 0. The equation is false! Notice what happened here. The substitution method led us to an equation that is always false. That is the algebraic signal that the original system of equations was inconsistent. There is no combination of X and Y that can solve both equations at the same time. A Dependent System One more unusual case is possible for a system of linear equations. We ordinarily do not expect more than one point of intersection because two lines should intersect in a single point. But sometimes a system of two linear equations in two unknowns can intersect in an entire line. How could this happen? We saw in Chapter 2 that there are many ways to write an equation for a particular line. Such equations are all equivalent, even though they differ in appearance. Here is an example of a pair of equivalent linear equations: A. Y = <->0.5X + 4, B. X – 2Y = <->8. The equations appear different, but in this example the second equation was constructed by multiplying both sides of equation A by <->2 and then adding X to both sides. In short, these equations boil down to the same relationship between Y and X. Such a system is called dependent because one of the equations can be found simply by rewriting the other equation using the rules of algebra. These two equations represent only one distinct relationship between the variables X and Y. Because a linear equation can be written in different ways, it is possible to encounter a system of two linear equations that turns out in reality to be two different ways of writing the same equation! When we graph the two equations, they will appear as one and the same line as in Figure 3.11. {INSERT FIG. 3.11} What happens algebraically if you try to use the substitution method on a dependent system? Let’s try to solve the system above using the substitution method and find out. Example 12 A Dependent System Try to solve the system: A. Y = <->0.5X + 4, B. X – 2Y = <->8. Chapter 3 Systems of Linear Equations and Inequalities 22 Solution The top equation is already in slope-intercept form with the variable Y isolated, so we can substitute (<->0.5X + 4) in the place of Y in the bottom equation to get X – 2(<->0.5X + 4) = <->8. Expanding the parentheses by the distributive rule gives us the equation, X – 2(<->0.5)X – 8 = <->8, which reduces to X – X – 8 = <->8 and still further to <->8 = <->8. The bottom equation says that <->8 must equal <->8, which is always true no matter what values of X and Y we might have in mind. The X and Y have fallen out of that equation. The emergence of an equation that is always true is the algebraic signal that our original system of equations was dependent. Therefore the solution set is the set of all (X, Y) pairs that satisfy the first equation, Y = <->0.5X + 4. SUMMARY This section introduced an important new type of mathematical problem: solving for more than one variable at once when you have more than one equation relating the variables. This is called solving a system of equations. The equilibrium price and quantity in a market is a problem that requires solving a system of two equations. The problem of determining the effectiveness of advertising, the efficiency of an employee, and the “prices” of buildings and land can often be formulated in terms of a system of equations. The technique for solving a system of two equations and two variables can be either graphical or algebraic. Graphically, the solution to the system of equations is the point where the linear graphs of the equations intersect. A graphical solution is at best an eyeballed guess. The a more precise solution can be calculated algebraically by the substitution method in three steps: 1. Isolate one variable in one equation. 2..Substitute the formula for that variable into the other equation and solve that equation for the second variable. 3. Substitute that value for the second variable into the equation for the first variable and solve to get a value for the first variable. This method works almost all the time. There are two special cases where the substitution method runs into trouble, and in both cases, it is because there is no unique solution to the system of equations. The first, the inconsistent system, occurs when the two equations have the same slope but different intercepts, meaning that they never intersect at all. The algebraic signal of that problem is that the substitution method results in an equation that is always false. The other problem, the dependent system, occurs when the two equations are just different ways to write one equation. Then the solution is the entire line of the single equation. This is indicated when the substitution method results in an equation that is always true. Chapter 3 Systems of Linear Equations and Inequalities 23 EXERCISES Skill Builders For each of the following systems of equations, sketch a graph of the equations, and estimate the solution by looking for the point of intersection. Then solve the system of equations algebraically using the substitution method. 1. Q + P = 100, Q = 0.75P – 1.5 2. Y = 4X + 10, Y = 50 – 6X 3. 3X + 4Y = 270, 5X + 2Y = 100 4. 0.44A + 1.07B = 29.51, 0.80A + .55B = 24.35 2 X + Y = 13, X + 2Y = 11 A + B = 17, 6. 2A − B = 7 3P − 4Q = 8, 7. P + 2Q = 3.5 X + Y = −2, 8. X −Y = 6 R = S + 150, 9. R + S = 200 2 X + 3Y = 30, 10. 8 X − Y = 16 X Y =1 + 11. 30 20 Y = X + 5 5. 0.1X + 0.3Y = 11, 0.2 X + 0.5Y = 19 12. Solve the following systems of equations using the substitution method, and check your answer by substituting the solution into both of the original equations. 13. 3X – 6Y = 9, 2X + Y = 11 14. P = 200 – 5Q, P = 60 + 2Q 15. <->5A + 8B = <->10, 15A + 2B = 160 16. Y – 9 = 2(X – 5, Y – 1 = 3(X – 2) Chapter 3 Systems of Linear Equations and Inequalities 24 7 X + 3Y = 29, X +Y = 7 X + 2Y = 8.5, 18. 2 X − Y = 4.5 5 X − 6Y = 10, 19. 3 X + Y = 52 X + Y = 12, 20. X − 2Y = 18 5 X − 3Y = 23, 21. 4 X − Y = 24 V = 80 − 2U , 22. V = 2U + 40 4 P + T = 3, 23. 3P − 0.5T = 6 Y − 3 = 2( X − 4), 24. Y − 7 = 3( X − 5) 17. Use the substitution method to find out if the following systems of equations are either inconsistent or dependent. 25. 3X + 5Y = 20 45Y = <->27X + 160 26. <->2X + 5Y = 40 Y = 0.4X + 8 C − 2 D = 40, C − 3 = 2( D + 18) + 1 2 X + Y = 15, 28. Y = 20 − 2 X 3 X − 2Y = 8, 29. Y = 1.5 X − 4 12 X + 16Y = 120, 30. Y = 9.5 X − 0.75 27. Applications 31. Temperature. The centigrade, or Celsius, scale of temperature sets 0 degrees to be the freezing point of water and 100 degrees to be the boiling point at sea-level pressure. The Fahrenheit measure (F) is related to the centigrade measure (C) by the equation F = 32 + 1.8C. Find the temperature at which the measurement in degrees centigrade is the same as the temperature measured in degrees Fahrenheit. (Hint: What equation says that the temperatures as measured by each scale are the same?) Chapter 3 Systems of Linear Equations and Inequalities 25 32. Demand and supply. The quantity (Q) of pumpkins that would be demanded by consumers at Halloween time in a particular town is estimated to depend on the price (P) of a pumpkin by the equation Q = 25,000 – 2000P. Farmers in the region are not willing to plant pumpkins unless they can expect to earn a profit on them, but if they foresee a high price of pumpkins they will plant more acres of pumpkins and produce more. The farmers’ decisions to produce a supply of pumpkins is captured by the equation Q = <->10,000 + 10,000P. A. Draw a graph of the supply and demand “curves” in the pumpkin market. B. At what price will the quantity demanded by the consumers be equal to the amount that the farmers would be willing to supply? 33. Estimating productivity. The company Crystal Giftware produces two types of product. The Sunview is a decorated disk of glass that hangs in a window frame. The Lamp view is a prism of decorated rectangular glass panels that are soldered together and surround a small light bulb that illuminates the design. Manufacturing employees work a 40-hour week. The Marketing Department gives them a production request each week, detailing the number of Sunviews and Lampviews that need to be produced to keep up with current and expected demand. Production records indicate that when 6 employees were working, in 1 week they produced 160 Sunviews and 80 Lampviews. In another week, they produced 70 Sunviews and 105 Lampviews. How many labor hours does it take to produce 1 Sunview? One Lampview? 34. Lakefront real estate. Along Lake Michigan in Illinois and Wisconsin, real estate sells at a premium if it is right on the lake. For property that has no houses on it, it seems reasonable to assume that the two most important factors in the cost of the lot are its acreage and its frontage on the lake. A real estate agent has just provided you with data on the costs at which several properties sold recently in the area where you want to invest in some land. Property Acreage Frontage on (A, lake (F, feet) Cost ($) acres) X 6.3 300 115,000 Y 24.0 500 300,000 Z 6.1 550 165,000 The real estate agent is offering you a small lot that has 2.0 acres and 150 feet of lake frontage, for a cost of $40,000. Is this a good value? A. The linear model says cost = cA + dF, where c is the price per acre and d is the price per foot of lake frontage. Using the linear model, estimate the individual prices on acreage and on frontage from the data above for properties X and Y. B. Test your model by seeing how well it predicts the cost of property Z. C. Based on the stated cost, should you buy the lot with 2.0 acres and 150 feet of frontage if it is for sale at $40,000? 35. Advertising response. Eva markets her new cookbook, Root Vegetable Dishes from Denmark, by taking out small advertisements in two magazines, Veggie Life and Vegetarian Times. The ads in VL cost $100 each, and the ads in VT cost $175 each. People who want to buy the book call a toll-free number and place the order with a “fulfillment service” firm that takes their credit card number and ships the cookbook. For 2 months, Eva has placed advertisements in both magazines, more in VT than in VL because VT has the larger circulation. The number of ads and the resulting number of orders are shown in the table. Month of Ads January February Veggie Life (number of ads) 1 1 Vegetarian Times (number. of ads) Book Orders 2 2 170 174 A. Eva wants to determine how many responses are coming from each magazine so she can see how much response she is getting per dollar of advertising spent at each place. Does the information in this table help her? If it Chapter 3 Systems of Linear Equations and Inequalities 26 does, what advice can you give about the best place to advertise? If it doesn’t help her, explain why not. B. In March, Eva altered her advertising strategy and took out 2 ads in VL and only 1 in VT. The result was 160 orders. Using this information with the February data, what do you calculate to be the numbers of orders per ad in VL and in VT? C. Which magazine had the higher ratio of sales per advertisement to cost per advertisement? 36. Diet planning. Suppose that the only items left in the college cafeteria are enchiladas and pecan pie. The dietary ingredients of these foods and minimum requirements are as listed in the table below. What combination of these foods would exactly fulfill an 18-year-old female’s daily requirements for iron and vitamin A? (Hint: Your answer may seem odd, even though your calculations are correct. Check your solution.) Enchilada Minimum daily Dietary ingredient (1 serving, 0.5 lb.) Pecan Pie (1/6 pie) requirement Iron (mg) 3.3 mg/serving 4.6 mg/slice 15 mg. Vitamin A (IUs) 2720 IU/serving 220 IU/slice 800 IU 37. Library usage. Ever since St. Catherine’s College began their new M.A. in Education program, the usage of books in the library seems to have increased. However, the library does not ask for a student’s status (graduate or undergraduate) when books are checked out, so it is hard to say if the increased usage is due to the graduate students. The head librarian at St. Catherine’s had the idea to compare usage at her institution to the library usage at St. Mary’s College across town. St. Mary’s has a larger enrollment, but they have an even larger percentage of graduate students due to their M.A. programmes in education, psychology, and management. However, the librarian believes that the library usage per student of each type, graduate and undergraduate, will be the same in each college. The enrollment figures and library usage for the two colleges are given at right. Undergraduate Graduate Library books College students students used per year St. 1,100 100 33,500 Catherine’s St. Mary’s 1,800 400 74,000 A. What does this information tell you about the number of books used per year by the typical undergraduate vs. that of the typical graduate student? B. Should St. Catherine’s consider charging a higher library fee to the graduate students? 38. A light money box. At the county fair, adult admission cost $4 and children’s admission was $1.50. The turnstile at the entry counted a total of 400 people attending the fair, but the booth attendants did not issue separate paper tickets for adults and children, as is often done elsewhere. At the end of the day, there was $900 in the money box. That greatly annoyed the county commissioner, who thought that one of the booth attendants must have been taking money out of the box. The commissioner was overheard to say, “There should have been a lot more adult fees. I was at the fair most of the day, and I would say there were 3 kids for every 2 adults, maybe 2 kids per adult at the most, but not more.” Determine if the commissioner is right about the possible theft of fair revenues and decide whether an investigation of the employees is warranted. (Hint: Does the number of adults and children implied by the revenue data fit the commissioner’s estimates of the ratio of kids to adults?) 39. Travel speeds. An aircraft flying between Des Moines and St. Louis covers a distance of 200 miles each way. On the trip toward St. Louis, the aircraft faces a headwind and takes 2 hours to make the trip. Returning to Des Moines, the trip takes only 1.5 hours. What was the aircraft’s airspeed, and what was the speed of the wind? (Hint: The ground Chapter 3 Systems of Linear Equations and Inequalities 27 speed (known distance/time) is equal to the airspeed minus the wind speed going into the wind, but it is airspeed plus wind speed going with the wind.) 40. Canoe trip. Scott and Eva paddled their canoe on the Sudbury River one afternoon to look at birds and wildlife. They started upriver and turned around after 45 minutes. The return trip took 30 minutes, and as they passed under Lee’s Bridge, Scott noticed that the current was flowing at about 8 inches per second. How many miles did they paddle? (Hint: Convert the current speed to miles per hour and look for two equations in two unknowns that say, “timespeed = distance.”) 41. Blend design. Two types of coffee beans, Brazilian and Ghanaian, are used to make a blend of ground and roasted coffee. Suppose that Brazilian beans cost $0.80 per kilogram and Ghanaian beans cost $0.45. What weight of Brazilian beans and what weight of Ghanaian beans should be used in 1 kilogram of the blend to make the overall cost of the blend equal to $0.60 per kilogram? (Hint: There are two equations in this problem. One describes the cost of the blend, and the other describes the weight of the blend.) 42. Employee compensation plan. Mary and Tabatha are partners in a small consulting firm, Employee Development Associates (EDA), that offers a seminar called “Exemplary Customer Service” to business firms. When they started their company, both Mary and Tabatha spent most of their time on marketing the seminar to the training directors of potential client firms. After several months, they had developed some regular customers and then started to spend more of their time on teaching the seminars. About 40% of their revenue goes to cover office expenses, and the other 60% can be divided between the owners. Mary is now expecting her first child. She wants to cut back her hours of work after the baby comes, and she might withdraw altogether from EDA for a while. Tabatha will continue to work full-time. This raises the question of how to divide the 60% of their revenues between them, particularly because of all of the work that Mary put in on marketing to develop clients who are now repeat customers and who will generate teaching revenues for EDA even while Mary is on leave. Their seminar involves 10 hours of instruction and usually has 15 participants for whom the client firm pays a total of $800 per seminar. From past experience, Mary and Tabatha agree that they have had to spend about 40 hours in sales efforts to land a contract for a course. They want their compensation system to give the same rate of hourly pay to the marketing and the teaching activities. What percentage of the $800 seminar fee should go to the person who teaches it, and what percentage should go to the person who landed the contract assuming: A. Tabatha teaches the seminar and Mary landed the contract? B. Tabatha teaches the seminar and both partners contributed equally to landing the contract? 43. Compensation design. Oakley Industries currently pays its 20 salespeople a flat salary of $2,000 per month rather than a commission on the sales that they make. As a result of this policy, the company has been losing its good salespeople and attracting a lot of mediocre ones. The sales manager is thinking about switching to a two-option commission system that has the same total cost as before because he thinks that’s a way to compensate both types of salesperson fairly. Salespeople who choose the first option would get a $750 monthly salary plus a commission equal to X% of their sales. This would be attractive to the ordinary salespeople, who tend to sell about $10,000 of product per month. The salespeople who choose the second option would earn no monthly salary but get Y percent of their sales. This policy has to be more attractive to the hot salespeople, the type who can sell $25,000 per month. The sales manager knows that not everyone can be a hot salesperson, but he does think that he can recruit about 12 ordinary salespeople and 8 hot ones. A. What commission rates X and Y (expressed as a percentage of sales) should he offer under each option so that his total monthly sales person compensation is the same as it is now, and yet enables a strong salesperson to earn $3,000 per month? Chapter 3 Systems of Linear Equations and Inequalities 28 B. Under this arrangement will an ordinary salesperson who takes the first option earn more or less than they are earning per month under the present system? Would an ordinary salesperson rather take the second option? Would a hot salesperson prefer the first option? Graphing Calculator Exploration All of the above exercises can be solved algebraically and then checked using a graphing calculator. But the following question requires repeated solution of the problem under different assumptions. This will be easiest to solve using a graphing calculator that can produce a solution after each change of parameters. 44. Criminal investigation. A bystander at the time of a bank robbery thought she saw the thieves jump into a Royal Taxicab in making their escape, but she could not be sure of the name “Royal.” Royal, which operates within the limits of a city 75 miles away by interstate highway, uses only a specific brand of car whose gas mileage on highway and city roads is 35 and 21 miles per gallon (mpg), respectively. One taxi was of unknown location at the time of the robbery. Royal Taxi records show that this car’s tank had been filled that morning. By evening, the 12-gal tank registered 1/4 full, and the mileage was 250 more than what Royal’s fleet manager had recorded for that car the day before. In similar legal cases where a car brand’s gasoline mileage data was used, the courts allowed an error of ±5 mpg for highway and ±4 mpg for city driving. Might the taxi driver be innocent? (Hint: You’ll have to solve this problem 4 times, once for each combination of the extreme values for the mileage data.) Group Exercise 45. Inconsistent systems of more than two equations. Crystal Giftware Corp. produces two products, the Sunview and the Lampview. Their 6 manufacturing employees each work a regular 40-hour week, unless they are absent due to family responsibilities or illness. Suppose that the production data indicate the following. Week Regular hours worked Overtime hours Lampviews worked produced Sunviews produced 1 240 0 80 150 2 220 0 95 100 3 240 20 95 150 4 235 10 90 140 A. Each member of your group should pick a different pair of weeks. Use the data for that pair of weeks to calculate how many labor hours are needed per Sunview and per Lampview. Do you get a solution? B. Compare all group members’ answers from part A. Do you get the same answers or different ones? If you get different answers, then the system of all four equations, taken together, can be called inconsistent. Think of a few reasons why you might be getting different answers. (Think about the real production process that is summarized in the table of data given here.) C. Having only the data in this table, what do you recommend as a “best” estimate of the number of labor hours that go into each Sunview and each Lampview? D. In part C above, you answered a statistical question. We will not see much statistics in this algebra course, but it is useful to know that statistics is the topic of mathematical study where precisely this type of issue is addressed. Using algebra, you will always get an answer to the question about labor hours for two products when you have two independent observations. The answer comes straight from the technique of solving a consistent, independent system of equations. However, when you have more than two observations and still only two variables, mathematically you have an “inconsistent system.” That is, there typically are no two labor- Chapter 3 Systems of Linear Equations and Inequalities 29 hour coefficients that can solve all of the (here four) equations simultaneously. So you have to settle for an estimate of the labor-hour coefficients that is a best approximation to the true value. Statistics offers ways to think about what is a “best” approximation and offers techniques to calculate that approximation. Discuss these points in your group, and make sure that everyone understands (1) the “problem” created by having more equations than variables, (2) the meaning of an “inconsistent system” of equations, and (3) what the field of statistics has to offer toward a solution of this problem. 3.2 Systems of Linear Inequalities Learning Objectives Graph a linear inequality. Write an inequality given a verbal description. Translate a table of data into a system of linear inequalities. Find the solution set of a system of inequalities graphically. Perform algebraic operations on, and invert, an inequality. Spotlight on Dietary Planning In the Spotlight that introduced the previous section, we looked at the problem of finding a menu that fulfills specific dietary requirements. Our example was of a diet composed entirely of spinach and potatoes. (The algebraic technique of that section works best with systems of only two variables, so we could use only two foods in that problem.) By solving a system of two equations in two unknowns, we found a menu that would fulfill an 18-year-old female’s requirements for 60 milligrams (mg) of vitamin C daily and 15 mg. of iron. {INSERT Photo 3.2} The fact is, however, that there are many more than two dietary requirements that a menu must fulfill. There is a minimum requirement for calories, protein, and various other vitamins and minerals. A diet cannot be designed to serve exactly the minimum requirements of all these. The diet must provide at least the minimum requirement. And for some ingredients, like sodium and calorie content, the diet may have to fit within a maximum allowance. That is, the diet would have to supply at most a certain amount. In the previous section of this chapter, we studied the solution to a system of linear equations. In this section, we will study problems that are expressed not as linear equations but as linear inequalities. A linear equation, such as Y = 3X + 6, is graphed as a line in the X-Y plane. An inequality is a relation between two variables in which the “=” sign of a linear equation is replaced by one of the inequality symbols: < (“less than”), > (“greater than”), ≤ (“less than or equal to”), or ≥ (“greater than or equal to”). So, given a linear equation in the variables X and Y, such as X + 3Y = 12, the corresponding linear inequality could be any of the following: X + 3Y < 12, X + 3Y ≤ 12, X + 3Y > 12, or Chapter 3 Systems of Linear Equations and Inequalities 30 X + 3Y ≥ 12. Inequalities that use the symbols “<” or “>” are called strict inequalities. They say that a quantity must be strictly less than or strictly greater than another quantity. The corresponding inequalities using the symbols “≤” and “≥” are said to be not strict, because they allow the two quantities to be equal. In addition to the application of linear inequalities in dietary planning, there are many important applications in management science, where problems of production and operations management are often expressed in terms of limitations on managers’ decision variables. Systems of linear inequalities have especially important applications in the management of factories in which several products are manufactured. In those applications, the inequalities define the limitations under which a production system must be operated. SOLUTION SETS FOR LINEAR INEQUALITIES The previous sections of this chapter have shown that the solution to a linear equation in two variables is a line, and the solution to a system of two linear equations in two variables is (usually) the single point where the two lines intersect. Single inequalities and systems of inequalities, in contrast, will usually have many solutions, comprising a region in the X-Y plane. It will be our task in this section to describe the solution set either algebraically or graphically. The solution to one linear inequality is a half-plane in the X-Y coordinate system that is specified by a boundary line, as the following example shows. Example 1 Solution Set of a Linear Inequality Find the solution set of Y > 3X + 6 graphically. Solution The solution procedure has two steps. 1. Graph the boundary of the given inequality. In our example, the boundary is the graph of the equation Y = 3X + 6. If the inequality is a strict inequality (strictly < or >), as in this example, draw the boundary as a dashed line. If it is not a strict inequality (either ≤ or ≥ ), draw a solid line. The equation Y = 3X + 6 is in slope-intercept form, so we immediately recognize that Y = 6 is the vertical intercept. The horizontal intercept comes from setting Y = 0 and solving, thus we find X = <->2. Since Y > 3X + 6 is a strict inequality, we draw a dashed line between these two points, as in Figure 3.12. {INSERT Fig. 3.12} 2. Shade in the solution region. To determine which side of the boundary line is the solution, pick any point in the plane but not on the boundary line as a test point and substitute the coordinates of that point into the inequality. • If the resulting inequality is “true,” then the half-plane containing that point is the solution. • If the inequality is “false,” then the other half-plane is the solution. We’ll use (0, 0) as our test point here. Substitute X = 0 and Y = 0 into the inequality Y > 3X + 6: 0 > 3(0) + 6 ? Chapter 3 Systems of Linear Equations and Inequalities 31 0 > 6; false. So the half-plane for the solution is the side of the boundary that does not contain our test point (0, 0). We shade the side of the boundary that represents the solution, as shown in Figure 3.13. {INSERT FIG. 3.13 Make a Note The points (x, y) in the two-dimensional coordinate plane that satisfy the equality form a boundary line for the solution to the inequality. The boundary of the inequality is graphed as a line in the plane. The points (x, y) that satisfy the linear inequality all lie on one side of the boundary line. INEQUALITIES AS “CONSTRAINTS” In common English, a constraint is a boundary or limit of some kind. For example, a student’s lifestyle may be constrained by the amount of pocket money they earn in a part-time job. When people talk about constraints, they often use words like “limitation” or “allowance,” and they may refer to a “ceiling” or “floor” value to describe the way the constraint affects the range of choices available to them. When working with, and writing, inequalities, these English words signal the type of constraint, meaning whether it is of the ">" or "<" variety. We’ll see how these words are used in the examples that follow. Think about This For each of the following words, think of an example where the word is used to describe an inequality. What is the direction of the inequality in your example? If you think the word could imply either direction, give an example of each for that word. allowance requirement limitation ceiling floor Maximum Allowance Constraints (<, ≤) Example 2 Dietary Planning Consider the problem of creating a single meal for a college student who eats only enchiladas and pecan pie. The following table shows the content of 1 serving of enchiladas (1/2 pound) and a onesixth portion of pecan pie in relation to minimum or maximum requirements. (The maximum requirements on calories and sodium content have been made up for this example.) Dietary ingredients Enchilada Pecan pie (1/6 piece) Daily requirement “Nutritive Value of Foods,” Home and Garden Bulletin, No. 72 (Washington, DC: US Government Printing Office, 2002), , <http://www.nal.usda.gov/fnic/foodcomp/Data/HG72/hg72.html> Chapter 3 Systems of Linear Equations and Inequalities Calories Calories Protein (g) Iron (mg) Sodium (mg) 235 235 20 3.3 1332 32 575 575 7 4.6 305 1,500 min. 2,500 max. 46 min. 15 min. 4,000 max. Each enchilada provides 235 calories, and each slice of pecan pie has 575 calories. The student must decide how many servings of each to take. Write and plot an inequality that expresses the following constraint, The number of calories provided by E servings of enchiladas, plus the number of calories provided by P slices of pecan pie, must not exceed 2,500. Solution A student who eats a number E of enchiladas will take in 235E calories. Likewise, by eating a total of P slices of pecan pie they will take in 575P calories. The diet expresses a maximum allowance on calories, an upper bound or limit. So we must write this English statement mathematically as a formula (for the calories) that is required to be less than the number 2,500. This calls for the “<” inequality, 235E + 575P < 2,500 (calories). The boundary equation for this inequality is 235E + 575P = 2,500. Its intercepts are 2, 500 2, 500 ≈ 10.6 for E and ≈ 4.3 for P. The inequality is not strict, so we draw the line 235 575 solid, as if to “shade in” the (good) points on the line. As a test point, we use (0, 0) and see that it results in the true inequality 0 < 2,500. Thus, the origin is part of the solution region to be shaded in. (See Fig. 3.14.) {INSERT FIG. 3.14} By a similar logic you can construct a maximum allowance constraint for the sodium content of the meal in the following question. Practice Each enchilada has 1,332 mg of sodium, and each slice of pecan pie has 304 mg of sodium. Assuming that the student should limit their sodium intake to 2,000 mg in this meal, write out an English sentence for the maximum allowance on sodium, and then write it as an inequality. Answer: 1332E + 304P < 2,000 (mg sodium) Minimum Requirements Constraints (>, ≥) In a dietary context, there may be minimum requirements for the intake of essential vitamins, minerals, and calories. These minimum requirements constraints all appear as the “>” or “>” variety. Example 3 Dietary Planning Suppose that in terms of protein, a diet must satisfy a constraint expressed in English as The number of grams of protein provided by E servings of enchiladas, plus the number of grams of protein provided by P slices of pecan pie, must be at least 46. Chapter 3 Systems of Linear Equations and Inequalities 33 Food research shows that enchiladas typically have 20 grams of protein per serving, and pecan pie has 7 grams of protein per slice. Write and plot an inequality that expresses the protein constraint. Solution We formulate the protein requirement as a “greater than or equal to” inequality, 20E + 7P > 46 (grams of protein). The inequality is not strict, so the boundary line is part of the solution. We therefore draw the boundary as a solid line, as shown in Figure 3.15. The test point (0, 0) does not satisfy the inequality, so the solution region is on the opposite side of the boundary line. {INSERT FIG. 3.15} Figure 3.15 shows that, in terms of protein, the daily diet should consist of at least a little more than 6 slices of pecan pie along with no enchilada, or at least a little more than 2-1/2 servings of enchilada along with no pecan pie, or a minimum given by various combinations in between. Practice Assume that a healthy diet requires at least 1,500 calories intake per day and 15 mg of iron. Each enchilada provides 3.3 mg iron, and each slice of pecan pie has 4.6 mg iron. Write out an English sentence for the minimum requirement on calories and another sentence for iron. Then write these mathematically as inequalities. Answers: 235E + 575P > 1,500 (calories) and 3.3E + 4.6P > 15 (mg. iron) SOLUTION TO A SYSTEM OF LINEAR INEQUALITIES We have seen that the solution to one inequality is a half-plane. What would be the solution to a system of inequalities? When we studied systems of linear equations, a solution was a point in the coordinate plane that simultaneously satisfied each of the two equations. The solution to a system of linear inequalities will be the set of all points that satisfy simultaneously each of the inequalities in the system. A system of three or more equations in two variables will usually have no solution at all because three lines do not ordinarily intersect all in one point. In contrast, a system of inequalities in two variables can have solutions even when there are more than two inequalities in the system. Example 4 Diet Planning with Many Constraints Let’s consider the minimum requirements for calories, protein, and iron in the college student’s diet problem. The table below reproduces a portion of the data shown in Example 1. Dietary ingredients Calories Protein (g) Iron (mg) Enchilada 235 20 3.3 Pecan pie (1/6 piece) 575 7 4.6 Daily requirement 1,500 minimum 46 minimum 15 minimum Chapter 3 Systems of Linear Equations and Inequalities 34 Solution Taken together, the requirements in this table can be expressed as a system of three inequalities. To those three, we should add the obvious constraints that neither the number of enchiladas eaten per day (E) nor the number of slices of pecan pie eaten per day (P) can be a negative number. The resulting system is the following: A. 235E + 575P > 1,500 (calories), B. 20E + 7P > 46 (grams of protein), C. 3.3E + 4.6P > 15 (mg. iron), D. E > 0 (enchiladas), E. P > 0 (pecan pie slices). To graph this system of inequalities, we will graph each inequality on the same coordinate system, shading in the half-plane that satisfies the inequality. The solution to the system is the set of points that satisfy all the inequalities simultaneously. This will be the region where all the shading coincides. On the same axes, we graph each inequality individually, following the steps described in Example 1: 1. Graph the boundary line, using a solid line for a non-strict inequality (). Label the line with a word that describes the constraint. 2. Pick a test point, such as (0,0), and substitute those coordinates into the inequality to determine which side of the boundary line is the solution region. Shade the correct region using a distinctive pattern for each inequality, such as horizontal, vertical, or angled hash marks. The nonnegativity constraints D and E are easy to graph. Their boundary lines are the horizontal and vertical axes. Together, these constraints shade in all the points above the horizontal axis and to the right of the vertical axis. The solution to the system is the region where all shading coincides. In this case, the solution is the part of Figure 3.16 that lies above and to the right of all the boundary lines. Draw a darker line along the boundary of that solution region to identify it clearly. {INSERT FIG. 3.16} A system of inequalities can combine minimum requirement constraints and maximum allowance constraints. In the example below, we consider a more complete system of inequalities for the diet problem, involving mixed constraints. Example 5 Diet Planning with a System of Mixed Constraints This table shows a portion of the table that was introduced with Example 2. It shows three minimum requirement constraints and two maximum allowance constraints. Write a set of inequalities that expresses these dietary requirements and graph the solution set for this system. Dietary Pecan pie Daily ingredients Enchilada (1/6 piece) Requirement Calories 235 575 1,500 minimum Calories 235 575 2,500 maximum Protein (g) 20 7 46 minimum Iron (mg) 3.3 4.6 15 minimum Sodium (mg) 1,332 305 4,000 Chapter 3 Systems of Linear Equations and Inequalities 35 maximum Solution We rewrite the information from this table in mathematical form as inequalities. The left side of each inequality is a formula written in terms of the quantity of enchiladas E and the number of pecan pie slices P eaten per day for the dietary ingredient (calories, protein, etc.). The right side of each inequality is the requirement or allowance for that dietary ingredient. The direction of the inequality (< or >) is determined by the type of constraint (minimum requirement or maximum allowance). A. 235E + 575P > 1,500 calories (minimum requirement), B. 235E + 575P < 2,500 calories (maximum allowance), C. 20E + 7P > 46 grams of protein (minute requirement), D. 3.3E + 4.6P > 15 mg iron (minimum requirement), E. 1332E + 304P < 4,000 mg sodium (maximum allowance). The food must be consumed in positive quantities, so we add the nonnegativity constraints, F. E > 0 enchiladas (nonnegativity), G. P > 0 pecan pie slices (nonnegativity). The graph in Figure 3.16 already has the minimum requirement inequalities A, C, and D. To those, we will add B and E to determine the solution set of this more complex system. You can set up these equations yourself by following the steps in the Practice box below. Practice 1. Find the vertical and horizontal intercepts of the two remaining inequalities, B. 235E + 575P < 2,500, E. 1,332E + 304P < 4,000. 2. Which kind of line, solid or dashed, should represent the boundaries of the inequalities B and E? 3. Does the point (0,0) test “true” or “false” when substituted into each inequality? Answers: 1. B, (0,100/23), vertical and (500/47, 0), horizontal. E, (0, 250/19), vertical and (1,000/333, 0), horizontal 2. Both should be represented by solid lines since they are not strict inequalities, i.e., each of them is “less than or equal to.” 3. The point (0,0) tests as “true” in both cases. So the half-plane containing (0,0) in both cases is the solution region. Figure 3.17 shows the final graph of the system. The solution set is represented by the region where all of the shading overlaps. The coordinates of any point in that region correspond to a feasible combination of enchiladas and pecan pie, given our diet constraints. {INSERT FIG. 3.17} Practice By examining Figure 3.17, determine whether each of the following diets does or does not satisfy all dietary requirements. A. 2.5 enchiladas and 4 slices of pie Chapter 3 Systems of Linear Equations and Inequalities 36 B. 4 slices of pie and 2 enchiladas C. 2 enchiladas and 3 slices of pie D. 4 slices of pie and 1 enchilada Answers: (A) no, (B) no, (C) yes, (D) too close to the boundary line to tell from the graph More than Two Variables? In reality, of course, a healthy diet must include more than just these two foods. But when a system of inequalities has more than two variables, it is not possible to graph the solution set in a plane. The solution set would be a region of three-dimensional space for a problem with three variables, a region in four-dimensional space for a problem with four variables, and so on. These kinds of solutions are much harder to visualize. However, our analysis of the twovariable case lays the groundwork for understanding problems that have any number of variables. ALGEBRAIC OPERATIONS ON INEQUALITIES An inequality in two variables defines a half-plane in the coordinate system. What algebraic operations on an inequality preserve its solution set, and which operations change it? Addition Rule for Inequalities Adding (or subtracting) any real number to both sides of an inequality does not change the solution set. For example, it is true that 3 < 5, so it is true that 3 + 2 < 5 + 2, which is to say 5 < 7. This rule can be applied to inequalities that represent a relationship between two variables. As an example, consider Y > 3X + 4. Its solution is the region above the boundary line Y = 3X + 4. Adding a number to both sides of this boundary-line equation does not change the boundary line or alter the solution region. Thus, another inequality having the same solution region is Y + 10 > (3X + 4) + 10. Multiplication Rule for Inequalities Multiplying (or dividing) an inequality on both sides by a positive number does not change the solution set. For example, it is true that 3 < 5, and it is likewise true that 2 3 < 25, meaning 6 < 10. As applied to inequalities with variables, consider Y > 3X + 4 again. Multiplying a positive number on both sides of the boundary-line equation does not change the boundary line or alter the solution region. Thus, another inequality with the same solution region would be 10Y > 10(3X + 4). However, multiplying or dividing an inequality by a negative number has the effect of reversing the direction of the inequality. We know that 3 < 5. But <->1 3 is not less than <>1 5. The inequality becomes reversed: <->3 > <->5. As applied to inequalities with variables, this means that when we multiply both sides of an inequality by a negative number, we must reverse the direction of the inequality to preserve its solution set. For example, Y > <->2X + 5 has the same solution set as: <->0.5Y < X – 2.5. Chapter 3 Systems of Linear Equations and Inequalities 37 Example 6 Inverting an Inequality The inequality P > <->5Q + 200 shows how to determine the permissible values of P given a number for Q. Invert that inequality, so that it shows how to find the permissible values of Q given a number for P. Solution Starting with the original inequality, P > <->5Q + 200, subtract 200 from both sides. This addition/subtraction operation does not change the direction of the inequality! The result is P – 200 > <->5Q. Now divide both sides of the inequality by <->5. This division by a negative number forces us to reverse the direction of the inequality, changing ">" to "<." 200 P – < Q, −5 −5 <->0.2P + 40 < Q. Now turn around the inequality, so it reads with Q on the left side, Q > <->0.2P + 40. SUMMARY An inequality describes a situation where one quantity is to be greater than, or greater than or equal to, or less than, or less than or equal to, another quantity. We saw that linear inequalities can express minimum requirements (as “>" or “≥” constraints) or maximum allowances or resource limitations (as “<“ or “≤” constraints). The solution to an inequality involving two variables is the set of all points in the plane that satisfy the inequality. These points will all lie to one side of the boundary line of the inequality, which is the graph of the equation that corresponds to the inequality. The solution to one inequality is a half-plane. The solution to a system of inequalities is the set of points in the coordinate plane that satisfy all the inequalities simultaneously. This region of points corresponds to the intersection of the half-planes that solve each of the individual inequalities, and it is bordered by the graphs of their boundary lines. Algebraic manipulation of inequalities is similar to the solution of equations. The rules for addition and the rule for multiplication by positive numbers are the same as for equations. However, one rule is different. When you multiply or divide an inequality by a negative number, you must change > to < or vice versa, which is called reversing the direction of the inequality. We may turn around an inequality, for example changing X > 2Y to 2Y < X. This is not an algebraic operation but simply a logical restatement of the inequality. Notice the difference between reversing the direction (due to multiplying by a negative number) and turning around (restating) an inequality. Chapter 3 Systems of Linear Equations and Inequalities 38 EXERCISES Skill Builders In the X-Y coordinate system, graph the solution set to each the following linear inequalities taken by itself. 1. Y < 3X + 10 2. Y > 14 – 2X 3. 3X + 5Y < 15 4. 3X + 5Y > 15 5. Y < 10 6. X > 5 7. 2X – Y > 8 8. –2X + 3Y < 20 9. Y – 6 > 2(X + 3) 10. 2Y < 3X + 5 11. Y 3 X + 10 12. Y −3 X + 10 13. Y −3 X + 10 14. Y + 2 X 6 15. Y + 2 X 6 16. 2 X 5 (This means “ 2 X and X 5 .”) 17. −4 Y −1 18. 2 X + 3Y 12 19. 2 X + 3Y 6 20. 6 2 X + 3Y 12 Graph the solution set for each of the following systems of inequalities. 21. 3X + 7Y ≤ 150 5X + 4Y ≤ 100 3X + 8Y ≤ 120 X≥0 Y≥0 22. X + Y ≥ 90 5X + 12Y ≥ 600 4X + 6Y ≥ 240 X≥0 Y≥0 23. 4X + 5Y ≥ 360 2X + 4Y ≥ 200 5X + 9Y ≤ 400 X≥0 Y≥0 24. 3X + 5Y ≤ 400 3X + 5Y ≥ 200 2X + 10Y ≥ 350 Chapter 3 Systems of Linear Equations and Inequalities 39 Y≥X X≥0 Y≥0 X 0, Y 0, 25. 40 X + 50Y 300, 10 X + 5Y 35 X 0, Y 0, 6 X + 8Y 480, 26. 5 X + 4Y 100, X 60, Y 40 X 0, Y 0, 27. 5 X + 2Y 20, 6 X + 5Y 60, 2 X + 5Y 40 X 0, Y 0, 28. X + Y 15, Y 0.5 X + 2, X + 3Y 30 Applications 29. Production constraint. A wood shop manufactures tables and chairs. Assume that every table requires 15 minutes on the lathe machine, and every chair requires 20 minutes. Write an inequality that says, “The time used for lathing T tables plus the time used for lathing C chairs must not exceed 10 hours each day.” Sketch the graph of this inequality. 30. Inventory constraint. A computer retailer sells computers and monitors. Every computer in inventory. Write an inequality that says, “The total inventory space that can be occupied by C computers plus the inventory space occupied by M monitors must not be more than 500 square meters.” Sketch the graph of this inequality. 31. Diet planning. Assume that every kilogram of rice provides 10,000 calories to a diet, and every kilo of beans provides 7,500 calories. Write an inequality that says a diet of rice and beans together must provide a minimum requirement of 2,000 calories per day. Sketch the graph of this inequality. 32. Blending coffee. The bitterness of a blend of coffee is a weighted average of the bitterness (as judged by expert tasters) of the types of coffee that are mixed in the blend. Suppose that the bitterness index of coffee type A is 3 and that of coffee type B is 7. Write an inequality that says the bitterness of a blend of A and B must not exceed the bitterness index Chapter 3 Systems of Linear Equations and Inequalities 40 value of 5.5. Sketch the graph of this inequality. (Hint: As variables, use the percentage of each type of coffee in the blend, and remember that those two percentages must sum to 100%.) 33. Budget constraint. Alice has a monthly budget of $40 for recreation. Suppose that she goes to a small rural college where the only kind of fun is going to a movie or eating out. A movie costs $5, and a dinner out costs $10. Write an inequality that shows what combinations of movies and dinners out per month she can afford and sketch its graph. 34. College admissions. The Admissions Office at State University has a simple rule for admitting applicants, which is based on the applicant’s high school grade-point average (GPA) and their score on the SAT exam. A student will be admitted if they satisfy any of the following three conditions: (1) their GPA is at least 3.5, or (2) their total SAT score (maximum 1,600) is at least 1,200, or (3) the sum of their SAT score plus 200 times their GPA is at least 1,750. Prepare a graph that shows which students will be admitted and which will not. 35. Loan screening. A local bank gives out small business loans after evaluating the loan applicant’s business plan and personal assets. The loan officer rates the business plan on a 010 scale and likewise rates the person’s assets on a 0-10 scale. The loan officer puts twice as much importance on the quality of the business plan as on the personal assets, so the officer forms a total score for the loan application by adding the assets rating to 2 times the business plan rating. A loan application that gets a total rating of 21 or higher will be funded. Write an inequality to describe the banker’s criterion for awarding small business loans and sketch a graph to show which applications will get funded and which ones will not. 36. Health care screening. A doctor can make a good guess about the cause of a child’s infection by taking the child’s temperature (T) and white blood cell count (W). A normal temperature is 37 to 37.5oC. A normal white cell count is 4,000 to 8,000 cells per cubic millimeter of blood. A bacterial infection is more likely than a viral infection if T > 39oC and W > 15,000 cells per mm3. Sketch a graph of the regions in the T-W plane that correspond to normal health, viral infection, and bacterial infection. For the domain of temperature on the horizontal axis, use values from 37oC (98.6oF) to 41oC (105.8oF). 37. Phase diagrams. Whether a chemical will be in a solid, liquid, or gaseous state depends on the temperature and pressure of its environment. The graph that shows which combinations of temperature (T) and pressure (P) results in which state of matter is called a phase diagram for that chemical. The phase diagram is commonly plotted on a coordinate system with T on the horizontal axis and P on the vertical axis. Three distinct regions in the TP plane correspond to the states of solid, liquid, and gas. As an example, carbon dioxide (CO2) is a gas at room temperature (25oC) and sea level (1 atmosphere pressure). But if its temperature is decreased or its pressure is increased, it can become a solid (dry ice) or exist in liquid form. In the case of carbon dioxide, the temperature at which it becomes a gas is, to a fair approximation, linearly related to the pressure. Specifically, carbon dioxide becomes a gas when T > 5.35 P – 83.4, with T measured in degrees centigrade and P measured in atmospheres. A. Over a domain of temperatures from <->100oC to <->0oC, sketch part of the phase diagram for CO2. That is, graph the region on the T-P coordinate grid where carbon dioxide is in gaseous form. (Hint: To draw it on a grid with P on the vertical axis, you may want to invert the inequality so that it reads P < . . . .) B. Carbon dioxide exists in liquid form only at pressures above 5 atmospheres. At those pressures, it will be a solid, rather than liquid, only if T < 0.8P – 60. Add this inequality to the previous one to complete your phase diagram for CO 2, and label the regions that correspond to solid, liquid, and gaseous states. Pressure can also be measured in pounds per square inch. Atmospheric pressure at sea level is 14.7 lb./sq. in. This is defined to be a pressure of “1 atmosphere.” Chapter 3 Systems of Linear Equations and Inequalities 41 C. At the ordinary sea-level pressure of 1 atmosphere, what temperature is required to form dry ice? 3.3 Characteristics of a System of Linear Inequalities Learning Objectives Determine the amount of slack in a constraint and evaluate its meaning. Determine when a constraint is redundant. Recognize when a system of inequalities is inconsistent. Understand the reality behind unbounded solution sets. Spotlight on Slack Time in the Admissions Office “This is extraordinary,” said Joanne Fennell as she returned from the campus post office one morning. “It looks like this week we’ll have 50 more applications for admission than last week, and I’d say that about a third of them are from foreign students.” “That’s both good news and bad news for me,” replied Corrine Early, the director of Admissions. “The more applications the better, but it takes a lot of our staff’s time to review them. I had to press our vice-president quite hard just to get three extra people last week to help with admissions processing. I was hoping those employees would give our office a bit of slack, but it now looks as though we’ll be back working at full throttle again.” “Perhaps not,” said Joanne. “Each American application takes about 2 hours to review, and each international application about 3.5 hours. I think those three people can handle the extra load, but I will have to sit down and do a few calculations to make sure.” {INSERT Photo 3.3} SLACK CONSTRAINTS AT A POINT Figure 3.18 shows two points within the solution region of a “minimum requirement” type of linear inequality. At point A, which is on the boundary of the solution region, the inequality is said to be binding, in other words, on the boundary line, the inequality is satisfied as an equality. Point B lies in the interior of the solution region. At that point, the inequality is said to be slack or “have slack.” This means that at point B, the right-hand side of the inequality is more than the minimum required. {INSERT Fig. 3.18} The concept of a constraint being binding or slack at a point applies just as well to a system of inequalities. If the point lies on the boundary of the solution set for the system, then at least one constraint will be binding, meaning that that particular constraint has no slack. At a point on the boundary of the solution where two constraint lines intersect, each of those constraints will be binding, and all other constraints will be slack. The analysis of slack constraints is very important in production and operations management. We shall see that when a constraint is slack at a point, it means that the righthand side of the constraint can be increased or decreased, and yet the point will remain in the solution region. Chapter 3 Systems of Linear Equations and Inequalities 42 Example 1 Slack in the Calorie and Protein Constraints Our solution to the diet problem was the region in the graph shown in Figure 3.19, including the boundary lines. {INSERT FIG. 3.19} Notice that 2 enchiladas and 3 slices of pecan pie is a feasible daily menu in this example. At this point, how much slack is there in the protein requirement, and how much in the calorie allowance? Solution We look first at the mathematical statement of the protein requirement, 20E + 7P ≥ 46 grams of protein. We substitute the values E = 2 and P = 3 into the left side of this constraint and find out how much slack remains in the requirement, 20(2) + 7(3) ≥ 46 grams of protein, 61 ≥ 46. Evidently, the slack in the protein constraint at the point (2, 3) is 61 – 46 = 15 grams of protein. This diet provides 15 grams of protein more than the minimum requirement. Put another way, the minimum requirement could be increased by 15 grams and this menu would still be feasible. Now, the calorie allowance says, 235E + 575P ≤ 2,500 calories. We use the values E = 2 and P = 3 in this inequality to find out how much the left side remains below the right side: 235(2) + 575(3) ≤ 2,500, 470 + 1725 ≤ 2,500, 2,195 ≤ 2,500. The slack in the protein constraint is 2,500 – 2,195 = 305 calories, meaning that the diet provides 305 calories less than the maximum allowance. Another way of looking at this is to say that the calorie allowance could be reduced by 305 calories, to 2,195 per day, and the diet of 2 enchiladas and 3 slices of pecan pie would still be feasible. Example 2 Slack in a Manpower Constraint Consider the limitation on the use of manpower in a wood shop to produce the quantities B and T of bookcases and tables, respectively. If each bookcase requires 3 hours of labor and each table uses 4 hours of labor, then the constraint that weekly production must not use more than 80 labor-hours can be expressed by the inequality 3B+ 4T ≤ 80 labor-hours per week. If this week’s production schedule calls for B = 12 bookcases and T = 10 tables, is there slack in the constraint? If so, how much? Solution Chapter 3 Systems of Linear Equations and Inequalities 43 The total labor requirement will be 3(12) + 4(10) = 76 hours, which is 4 hours less than the limitation in the constraint. So we would say that at the point (12, 0) in the B-T plane, the labor constraint has 4 hours of slack. (See Fig. 3.20.) The amount of slack in a constraint at any one point is the quantity by which we would have to change the right-hand side of the constraint in order to make the constraint binding at that point, as the following example demonstrates. {INSERT FIG. 3.20} Example 3 Slack in the Admissions Office The opening Spotlight for this section introduced the example of a school that had received 50 more applications than they were expecting that week. International applications require 3.5 hours of staff time to process, but American applications require only 2 hours. The office had received 3 new staff people, who each work 40-hour weeks. Can these people handle the additional load of the new applications? Solution The constraint on staff time, stated in words, is that the total number of hours needed to process the American applications plus the number of hours needed to process the foreign applications cannot exceed the total of 120 hours of staff time available from these three employees. If we write A for the number of American applications in the batch and F for the number of foreign applications, the constraint can be stated mathematically as 2A + 3.5F ≤ 120. The batch of 50 new applications appeared to be about one-third foreign and two-thirds American, meaning about 502/3 ≈ 33 Americans and 501/3 ≈ 17 foreign. Will that leave any slack in the staff-time constraint? Figure 3.21 shows a graph of this inequality together with the point (A,F) = (33, 17), which represents the needed manpower. The point seems to be outside the solution region, but it is too close to the boundary line to tell visually. {INSERT FIG. 3.21} Mathematically, we can determine precisely whether or not the point is in the solution set. We find out by substituting F = 17 and A = 33 into the inequality and testing the truth of that mathematical statement: 2(33) + 3.5(17) ≤ 120 66 + 59.5 ≤ 120 125.5 ≤ 120; False. It is not in the solution set. The conclusion is that 33 American applications and 17 foreign applications would more than use up the available time of the new staff. There will not be any slack personnel time in the Admissions Office this week! Practice 1. What is the amount of slack in the inequality 5X + 4Y < 100 (hours of machine time) at the point (X,Y) = (9,12)? 2. What is the amount of slack at the point (10,13)? What does this tell you about the slack in a constraint at a point that is outside the constraint’s solution set? Chapter 3 Systems of Linear Equations and Inequalities 44 Answers: (1) 7 hours of machine time. (2) The slack is negative at points outside the solution set of that constraint. REDUNDANT CONSTRAINTS In the course of solving realistic problems, it is not unusual to discover a new limitation constraint or a new requirement to be fulfilled. Adding the new constraint may change the set of points in the solution. Or, as you will see in the next example, it may not. When an inequality can be added to a system, or removed from the system, without changing the solution set, that inequality is called redundant. Example 4 Diet Problem Returning to our diet problem, assume that a minimum daily requirement of 1,000 IUs of vitamin A has been added to our list of constraints. Each enchilada provides 2,720 IUs of vitamin A, and each slice of pecan pie provides 220 IUs. Write an inequality that describes this constraint and decide if it’s redundant. Solution Express this constraint as 2,720E + 220P > 1,000 (IUs of vitamin A). The vertical and horizontal intercepts of this expression give an idea of where this constraint lies in the enchilada-pie plane. Set E = 0 and you see that the constraint can be met by 4.5 slices of pie. But setting P = 0 and solving for E tells us that 1,000 or about 220 1,000 or about 0.37 of an 2,720 enchilada alone will meet the constraint. The vitamin A constraint will have a boundary line that passes between these two intercepts, as shown in Figure 3.22. When this constraint is added to the set of minimum requirements and maximum allowances that was graphed in section 3.2, we see that the vitamin A constraint falls far away from the set of solution points. {INSERT FIG. 3.22} This tells us that adding or removing the vitamin A constraint from the other five constraints does not change the solution. The vitamin A constraint is therefore redundant. Another example is seen in the following system of inequalities. Example 5 Identifying Redundant Constraints Identify the redundant constraints, if any, in the following system: 1. 2X + 5Y 2. 7X + 4Y < 28, 3. 2X + 2Y < 12, 4. X ≥ 0, 5. Y ≥ 0. < 15, Chapter 3 Systems of Linear Equations and Inequalities 45 Solution When we plot these on the same coordinate system, we get the graph shown in Figure 3.23. {INSERT FIG. 3.23} The solution set is the region in the lower-left portion of the graph, not including the boundary lines because the inequalities are strict (<). Inequality 3 does not pass through this region, so its removal from the system would not alter the solution set. Another way to see this is that adding inequality 3 to the system does not restrict the solution set any more than what the inequalities 1, 2, 4, and 5 together do. Inequality 3 is redundant in this system. In contrast, the attempt to remove any of the other constraints, 1, 2, 4, or 5 would alter the region that represents the solution. If inequality 1 were removed, the region would then extend up as far as the line for inequality 3. If inequality 2 were removed, the region would then extend out to the right as far as the line for inequality 3. These facts show that inequalities 1 and 2 are not redundant within the entire system of inequalities. INCONSISTENT SYSTEMS OF INEQUALITIES In the diet example with pecan pie and enchiladas, the solution set was a small bounded region in the plane. There are other types of solution sets for systems of linear inequalities, and sometimes a solution may not exist at all. The graph of the solution to the following system appears in Figure 3.24: 1. 2X + 5Y ≤ 15, 2. 7X + 4Y ≤ 28. The figure also includes the inequalities X ≥ 0 and Y ≥ 0. {INSERT FIG. 3.24} Now consider a third inequality, 3. 2X + 2Y ≥ 12. When the graphical solution of inequality (3) is added to Figure 3.24, there appear to be two separate solution regions (Fig. 3.25). But the solution to a system of inequalities is the region that satisfies all inequalities simultaneously. The appearance of two distinct regions means that there is no point or region that satisfies inequalities 1, 2, and 3 simultaneously, along with inequalities X ≥ 0 and Y ≥ 0. {INSERT FIG. 3.25} Such a system of inequalities is called inconsistent and has no solution. UNBOUNDED SOLUTION SETS A system of inequalities can have a solution that is infinitely large. If we consider only the minimum dietary requirements in the diet problem, an infinitely large lunch would still meet all of the constraints. This is evident in Figure 3.26 from all of the shaded space (solution set) that goes infinitely far in the upper-right direction of the coordinate system. Only when we imposed a maximum allowance constraint, such as the one on sodium or on calories, did the solution set become a bounded region in the coordinate plane. Chapter 3 Systems of Linear Equations and Inequalities 46 {INSERT FIG. 3.26} Realistic problems in operations management always have bounded solution sets. For example, suppose you are using linear inequalities to describe the production of tables and chairs with limited amounts of time, labor, and raw materials. If you graph the inequalities and come up with an unbounded solution set, it is an indication that there is a numerical mistake somewhere in your system of inequalities or that some limitation constraint is missing entirely. No production system can produce an infinite amount of a product. For systems of inequalities that describe the requirements of a diet or of an animal feed, the solution set often is unbounded. But that does not mean that the dietitian in a hospital, school, or military base is going to prescribe a diet that gives each person an infinite amount of carrots. In practical applications of the diet problem, the linear inequalities are used as a backdrop for a problem of trying to find a diet that meets the requirements at lowest cost. In the search for a lowest-cost menu, a dietitian avoids the unbounded part of the solution region even though it is mathematically feasible as a diet. LINEAR PROGRAMMING The actual decisions faced by a production manager are made with some purpose or objective in mind, such as to maximize profits or to minimize costs, under the constraints imposed by the physical production system and other organizational requirements. The most important application of linear inequalities, which we have studied in this section, is to define the set of feasible choices in this type of decision problem. The technique that is often used to find the maximum-profit point or minimum-cost point among the feasible points is called linear programming. The problem of maximizing a linear objective function subject to a collection of linear inequality constraints is called a linear programme. There is an elegant solution to the standard linear programming problem, which works for any number of variables and any number of constraints when the inequalities have at least one solution. That solution procedure is called the simplex algorithm. The mathematical ideas behind it are an application of matrix algebra, which is studied in more advanced courses. The simplex algorithm does more than just locate the values of the decision variables that maximize or minimize the objective. In the process of calculating the solution, the simplex algorithm provides vital information about the effects of changing the numerical values of the constraints or of the decision variables. These insights provided by the simplex algorithm’s solution go beyond what we can see in a simple graphical solution. In fact, mathematical analysts use computers to calculate the complete solution to a linear programming problem in any realistic application to profit maximization or cost minimization. The subject of linear programming is so important to business decision making that the topic rightly deserves to be a major part of a separate course. Business majors will encounter the theory and practical applications of linear programming in a course on operations research or on “quantitative methods” in business. Understanding the theory of linear inequalities, as presented in this section, provides a basis for the study of linear programming. SUMMARY Systems of linear inequalities are a mathematical tool used in many important applications in management and the social and health sciences, particularly in problems of maximizing profit or minimizing cost. A system of linear inequalities is often used to describe the constraints within which a person must make decisions. In this section, we followed an example where linear inequalities define the set of menus (albeit a rather limited menu) that meet a whole variety of constraints on a person’s diet. When a solution to a system of linear inequalities exists, the solution set may be bounded within a finite region of possibilities, as in our diet planning example (Fig. 3.19), or the solution set may be unbounded (Fig. 3.26). Given any point in the solution region, it is Chapter 3 Systems of Linear Equations and Inequalities 47 possible to calculate the amount of slack in a constraint, which tells us how much the righthand side of the constraint can be changed before the constraint becomes binding. We noted that sometimes a constraint can be removed or added without altering the solution to the other inequalities. Such a constraint is redundant within the system of inequalities. A system of linear inequalities is inconsistent if there is no point in the coordinate plane that can satisfy all constraints simultaneously. EXERCISES Skill Builders For each of the following systems of inequalities: A. Graph the solution set. B. Determine if the system is inconsistent. C. Determine if the solution set is unbounded. D. Identify the redundant constraints, if any. [ED: Please take out tabs; not needed for comp]1. 5X + 4Y ≤ 100 3X + 8Y ≤ 120 X≥0 Y≥0 2. X + Y ≥ 90 5X + 12Y ≥ 600 4X + 6Y ≥ 240 X≥0 Y≥0 3. 4X + 5Y ≥ 360 2X + 4Y ≥ 200 5X + 9Y ≤ 400 X≥0 Y≥0 X 40 Y 50 4. X + Y 100 X 10 Y 20 7 X + 4Y 280 10 X + 12Y 600 5. X + Y 60 X 0 Y 0 3X + 7Y ≤ 150 Chapter 3 Systems of Linear Equations and Inequalities 48 30 X + 20Y 1, 200 40 X + 50Y 2, 000 6. Y 2 X + 10 X 0 Y 0 8 X + 7Y 280 X + 2Y 60 7. 2 X + 3Y 60 X 0 Y 0 8. In the following system of inequalities, which constraints, if any, are redundant, and which form part of the boundary of the solution set? A. B. C. D. E. 4X + 5Y < 360 2X + 4Y < 200 3X + 7Y > 300 5X + 4Y < 280 3X + 8Y > 240 9. In the following system of inequalities, which constraints, if any, are redundant, and which form part of the boundary of the solution set? Pick one point at the intersection of two boundary lines. Calculate the amount of slack in each nonredundant constraint at that point. 9 X + 11Y 495 2 X + 3Y 90 5 X + 6Y 150 (Note to compositor: this should have (A-E labels as above). Y 15 + X Y 6 X − 12 Applications 10. Publicity. The student association has allocated $600 toward publicity for the upcoming Halloween Fest. Ads in the school newspaper cost $15 per column-inch. Color posters can be printed and posted around campus for $1.25 each. A. Write an equation for the publicity budget constraint and sketch its graph. B. The Halloween Fest director thinks that to publicize the event properly, she will need to use 25 column-inches of advertising and distribute 80 posters. Is this plan feasible given her budget constraint? C. How much slack is there in the budget constraint if she follows the plan described in part B? 11. Production constraints. Assume that every unit of product A requires 7 minutes on the grinding machine and every unit of product B requires 12 minutes. A. Write an inequality that says, “The time for grinding A units plus the time used for grinding B units must not exceed 8 hours each day,” and sketch its graph. B. How much slack is there in the grinding constraint if the current production schedule calls for 18 units of product A and 25 units of product B to be manufactured today? Chapter 3 Systems of Linear Equations and Inequalities 49 12. Inventory constraints. Assume that every unit of the radios in inventory uses 1.6 square feet. of space, and every unit of clocks uses 0.7 square feet. A. Write an inequality that says, “The total inventory space that can be occupied by X units of radios plus the inventory space occupied by Y units of the clocks must not be more than 500 square feet.” B. How much slack remains in the inventory space constraint if the warehouse currently has 180 units of radios and 250 units of clocks? 13. Diet planning. Assume that every kilogram of dry rice provides 10,000 calories to a diet, and every kilogram of dry beans provides 7,500 calories. A. Write an inequality that says a diet of rice and beans together must provide a minimum requirement of 2,000 calories per day. B. Suppose that a person has already eaten 100 grams (0.10 kg) of dry rice and 80 grams (0.08 kg) of dry beans? Put a dot on your graph indicating this point. How much slack remains in the calorie constraint at that point? What is the meaning of negative slack? 14. Dietary constraints. Suppose that a diet consists entirely of enchiladas and pecan pie. In this problem, we consider the dietary constraints on fat. Each gram of fat, saturated or not, contributes 11 calories to one’s diet. The table shows how much fat is contained in each serving of enchiladas and pecan pie Content of foods Enchilada Pecan pie slice (230 grams) (1/6 pie, 138 grams) Fats (all kinds) 16 grams 32 grams Saturated fats 7.7 grams 4.7 grams Energy 235 calories 575 calories A. Write inequalities that express the following 1996 dietary guidelines of the U.S. Dept. of Health and Human Services. • No more than 30% of the calories in the daily diet should come from fats. • No more than 10% of the calories in the daily diet should come from saturated fats. B. Are either of these constraints redundant when considered along with the constraints used in the version of the diet problem shown in Figure 3.19? 15. Admissions screening. The Great Plains School of Journalism accepts students into its master’s programme based on their verbal and analytical scores on the Graduate Record Exam (GRE). The school considers verbal ability to be slightly more important than analytic skills. Applicants are automatically admitted if the verbal score is at least 600, the analytical score is at least 550, and the sum of the verbal and analytical scores is at least 1,230. A. Sketch a graph of the regions in the verbal-analytic plane that correspond to automatic admission. B. On your diagram, mark the point corresponding to an applicant who has a verbal score of 590 and an analytic score of 650. Will this person be automatically admitted? C. Continuing from part (b), how much error could there be in the applicant’s verbal score before it would change the Admissions committee’s decision? 16. Admissions screening. The International Institute of Business admits students into its MBA program based on their English-language proficiency as measured by the TOEFL exam and their aptitude for business study as measured by the GMAT. They do not consider undergraduate GPA because their students come from many countries with different grading systems. Students at IIB are automatically admitted if their TOEFL is at least 575, their GMAT is at least 550, and the sum of the TOEFL and GMAT is at least 1,200. Chapter 3 Systems of Linear Equations and Inequalities 50 A. Sketch a graph of the regions in the TOEFL-GMAT plane that correspond to automatic admission. B. On your diagram, mark the point corresponding to an applicant who has a verbal score of 625 TOEFL and a GMAT score of 580. Will this person be automatically admitted? C. Continuing from part (b), how much error could there be in the applicant’s TOEFL score before it would change the Admissions committee’s decision? 17. Health screening. When a doctor evaluates a patient’s risk of heart disease, she considers the patient’s age, gender, total cholesterol, and systolic blood pressure among other factors. Write C for the total cholesterol (mg/dL) and B for the blood pressure (mm Hg). A 50-year-old man will increase his risk of heart disease in the next 10 years by 10 percentage points if the measures of cholesterol and blood pressure satisfy the inequality 0.05C + 0.10B > 30.6. A. Sketch a graph of the regions in the C-B plane that correspond to less than 10% extra risk and more than 10% extra risk heart attack. For the cholesterol measure use values between 140 and 330, and for the blood pressure measure use values between 100 and 180. B. On your diagram, mark the point corresponding to the observation of a 50-year-old male with a total cholesterol of 220 and a systolic blood pressure of 140. Has this person added more than 10 percentage points to their chance of heart disease? C. How much would the patient have to increase or decrease their blood pressure in order to change the doctor’s assessment of the risk category that person falls into (that is, more than 10% extra risk or less than 10% extra)? 18. States of matter. Propane is used as a fuel for lamps, heaters, and stoves. At the sealevel atmospheric pressure (14.7 lbs. per square inch, or 1 atmosphere), propane changes from liquid to gas at <->44oC. When the pressure is 17 atmospheres, propane becomes gaseous at a temperature of 40oC. A. Write an inequality that specifies the combinations of temperature and pressure at which propane is in a gaseous state. B. What amount of pressure is necessary to make propane liquid at the freezing point of water, 0oC? 19. Production constraints. Finney’s Wood Shop produces only tables and chairs. Thomas Finney, the owner, has three resources used in production, namely, inventory space, assembly labor, and painting labor. His shop is unionized, so the assembly workers do not do painting and vice versa. The requirements of space and labor for one table and for one chair are shown in the table below, as is the available quantity of each resource for the next week. Unit production Per Per Quantity requirements table chair available Inventory space (sq. ft.) 16.00 2.00 1,200 Assembly labor (hrs./wk.) 0.75 0.60 120 Painting labor (hrs./wk.) 0.50 0.20 80 Finney wants to make plans for the production of tables and chairs next week. He already has on hand orders for 40 tables and 100 chairs, which must be produced next week, or else customers will be very unhappy. A. Write five inequalities that define the constraints on Finney’s production plan for the next week. Hint: Three inequalities will pertain to resource limitations, and two will pertain to production requirements. K. M. Anderson et al. "An Updated Coronary Risk Profile," Circulation 83 (1991): 356-362. Chapter 3 Systems of Linear Equations and Inequalities 51 B. The feasible combinations of tables and chairs constitute a region in the coordinate system shown in Figure 3.27. Label each line in the diagram with the name of the appropriate constraint and shade the solution region for each. C. Give an example of a feasible quantity of tables and chairs for next week’s production. {INSERT FIG. 3.27} Analysis 20. Exercises 15 and 16 show that in an admissions process it is possible to describe the “automatic admission” criteria by a system of linear inequalities. A. In principle, can the criteria for “automatic rejection” be described by a system of linear inequalities? B. If you think it is possible, make up such criteria that are consistent with the data in Exercise 15. 3.4 Solving Systems of Linear Equations with More than Two Unknowns: Elimination Method Learning Objectives Solve a system of linear equations using the elimination method. Mathematically model a problem given its background and a table of data. Recognize algebraically when a system is either inconsistent or dependent. Spotlight on Decisions about Advertising “We should have instructed the receptionist to ask callers how they heard about these new programs,” said Donald Martinez, Assistant Director of Admissions. He looked over a report of the number of inquiries that Metropolitan University in Chicago had received over the last few months about its new graduate certificate programs in management. “Then we would really know which of our advertisements are producing results and which are not worth the money.” Carol Haines, a business major who was working in the Admissions Office several hours a week, overheard Mr. Martinez. She turned to him and asked, “How much advertising have you been doing?” “More than $10,000 per month. Why?” “I actually meant to ask how many different magazines or newspapers you are advertising in,” added Carol. Mr. Martinez quickly rattled off a list of four of the most prominent papers in the Chicago region. “I see,” said Carol. “How many months have you been collecting data about the inquiries for your graduate certificate programs?” “August, September, October, and November,” replied Mr. Martinez. “I think I can tell you how many responses are coming from each ad,” said Carol with a smile. “You’re kidding. How can you do that if we didn’t ask the callers where they heard about us?” “Using a technique I learned in an algebra course a while back. If you can tell me how many ads you placed in each source during each month, August to November, and you have the total inquiry data for those months, it will take me about 15 minutes to set up the problem on our office computer and solve it.” Chapter 3 Systems of Linear Equations and Inequalities 52 “That would be great! I have a report on advertising costs and total inquiries right here. It should contain the information you are asking for.” replied Mr. Martinez. He gave Carol the following table. {INSERT Photo 3.3} Source: Fortune Chicago Chicago Magazine Crain’s Chicago Business Suburban papers Price/ad: $3,500 $3,700 $2,300 $2,400 August ads September ads October ads November. ads 1 0 0 1 1 1 0 1 1 0 0 1 Inquiries received 3 2 Total monthly cost $13,000 $10,800 2 2 $8,500 $10,600 300 440 520 445 In the previous sections of this chapter, we saw how to solve a system of two equations in two unknowns. However, some problems in business and society require the simultaneous solution of three or more equations. As long as the equations are linear, there is a straightforward method to find a solution to such a system. This technique is called the elimination method. We can use it to solve a system of equations where there is a unique solution (analogous to a pair of crossed lines in the two-variable case) and also to identify the unusual cases where there is no solution or an infinite number of solutions. To understand how Carol would solve Mr. Martinez’ problem, we will first look at the elimination method in the simple context of two variables. THE ELIMINATION METHOD WITH TWO VARIABLES The elimination method works on a system of linear equations that has been written in general form. Consider the following system of two equations in two unknowns: A. 2X + 3Y = 10, B. 4X – 8Y = <->15. The objective of the elimination method is to reduce this system to an equivalent system in the form X = _____, Y = _____, where there are only numbers on the right-hand sides. Another way of looking at this is to say that we want the first equation in the equivalent system to have a coefficient of 1 on the X and 0 on the Y, and vice versa for the second equation, 1X + 0Y = _____, 0X + 1Y = _____. This technique is called the “elimination” method because the objective is to “eliminate” all but one variable from each equation. The way to “eliminate” a variable is to reduce its coefficient to zero following specific rules that do not change the solution to the system of equations. Chapter 3 Systems of Linear Equations and Inequalities 53 How does this method work in practice? The rules of algebra permit us to multiply an equation, on both sides, by a nonzero constant. At first it may be a surprise, but the rule of adding equal quantities to both sides of an equation also permits us to add one equation to another equation—left side to left side and right side to right side—because both sides of the one equation are equal. The key to solving a system of equations by the elimination method is to choose a multiplication factor for one equation in such a way that it will eliminate the coefficient of a variable in another equation when the two equations are added. Example 1 Solving a System of Two Equations in Two Variables Using the elimination method, solve for X and Y in the following system: A. 2X + 3Y = 10, B. 4X – 8Y = <->15. Solution To eliminate the X variable from the bottom equation, we must add <->4X to the existing 4X, which will result in 0X. Multiply equation (a) by –2, <->2(2X + 3Y) = <->2(10), resulting in the equivalent equation <->4X – 6Y = <->20. Now add this to equation (b), left side to left side and right side to right side, and call the resulting equation B1,[comp: align on equal signs] <->4X – 6Y = <->20, 4X – 8Y = <->15, B1. 0X – 14Y = <->35. Recall that our objective in this process is to create a system of equations that has only 0’s and 1’s as coefficients. In particular, we want the bottom equation to have 0 as its X<−1 >coefficient and 1 as its Y-coefficient. We accomplish this by multiplying equation B1 by : 14 −1(−14 ) −1 −35 (B1) 0X + Y= −14 14 14 = (B2) 0X + 1Y = 2.5. Now we replace the original equation B by equation B2. The new system of equations has the same solution as the original system of equations, but now it appears in a simpler form as A. 2X + 3Y = 10 B2. 0X + 1Y = 2.5 If we were following the substitution method, at this point we would substitute the known Y value, 2.5, into equation A and solve for X. However, the full-blown elimination method must continue with two more steps. (You’ll see later in this section that the substitution method would be very messy if we had more than two variables, yet the elimination method would continue to be quite simple. Here, we are using the elimination method on a twovariable problem just to show how it works.) The Y in equation A has a coefficient of 3. Our goal is a Y coefficient of 0, so we multiply equation B2 by –3 Chapter 3 Systems of Linear Equations and Inequalities 54 <->3(b2) + (A) 0X – 3Y = <->7.5 2X + 3Y = 10 = (A1) 2X + 0Y = 2.5 To finish the elimination method, we must turn the coefficient of X in equation A1 into a 1 1. To do so, we multiply the entire equation A1, left and right, by , 2 1 1 (A1) (2X + 0Y) 2 2 = = (A2) = 1.25 1X + 0Y 1 (2.5) 2 The new system of equations, using A2 and B2, has the same solution as the original system, but the solution is now immediately evident: A2. 1X + 0Y = 1.25, X = 1.25, which reveals that B2. 0X + 1Y = 2.5, Y = 2.5. We check the solution (X,Y) = (1.25, 2.5) by substituting these values into each equation of the original system: equation A: 2X + 3Y = 10, 2(1.25) + 3(2.5) = 10, 2.5 + 7.5 = 10;✓ equation B: 4X – 8Y = <->15, 4(1.25) – 8(2.5) = <->15, 5 – 20 = <->15.✓ Make a Note The elimination method transforms the appearance of a system of linear equations in a way that does not alter the solution to the system. In its final form, the system has coefficients of 1 and 0, which enable us to “read” the solution directly. SOLVING EQUATIONS WITH THREE OR MORE VARIABLES The following is an example of a system of three equations in three variables: A. X + 4Y + 2Z = 29, B. X + 3Y + 3Z = 35, C. 3X + Y + 2Z = 33. Trying the Substitution Method on Three Variables Before we apply the elimination method to this problem, let’s see why the substitution method becomes difficult with three or more variables. The first step in the substitution method would be to isolate one variable in one of the equations. The variable X is isolated in equation A, so we isolate it by rewriting equation A as X = 29 – 4Y – 2Z. The next step in the substitution method would be to substitute this formula for X in the other two equations, which then have only two variables each, Chapter 3 Systems of Linear Equations and Inequalities B1. (29 – 4Y – 2Z) + 3Y + 3Z = 35, C1. 3(29 – 4Y – 2Z) + Y + 2Z = 33. 55 This is now a system of two equations in two variables, Y and Z. We would have to simplify it by collecting like terms, B1. <->4Y + 3Y – 2Z + 3Z = 35 – 29, C1. 3(<->4)Y + Y + 3(<->2)Z + 2Z = 33 – 329, and further simplifying to B1. <->1Y + 1Z = 6, C1. <->11Y + <->4Z = <->54. At this point, one could continue the substitution method, writing equation B1 as Z = 6 + Y and substituting into equation C1 to get an equation in Y alone, <->11Y + <->4(6 + Y) = <->54, <->11Y + <->24 – 4Y = <->54, <->15Y = <->30, Y = 2. With that value for Y in hand, the substitution method would continue by using that value in equation (B1) to find Z = 8, and then using the known values for Y and Z in equation A to find X = 5. The first substitution step above created a rather messy system of two equations in two unknowns, but we were still able to push forward with the substitution method. If the system had had four equations in four variables, the first substitution step would have resulted in a very messy system of three equations in three variables, which would be even more difficult to pursue under the substitution method. In contrast, the elimination method will be more straightforward to apply and can be generalized much more easily to problems with four or more variables. We’ll now outline the steps of that procedure. Overview of the Elimination Method The basic elements of the elimination method are the two operations we performed on equations in Example 1 that preserve the solution to the system. One such operation was to multiply an equation, left and right, by a nonzero real number. The other operation was to add the left side of one equation to the left side of another equation and correspondingly the right side to right side. These two operations are called row operations because they apply to entire rows (equations) in the system of equations. Consider again the system of three equations on which we started the substitution method before: A. X + 4Y + 2Z = 29, B. X + 3Y + 3Z = 35, C. 3X + Y + 2Z = 33. The goal of the full elimination method is to “eliminate” all coefficients except that of the X in the first equation, Y in the second equation, and Z in the third equation. That would result in a system that appears in diagonal form, having coefficients of 1 along the top-left-tobottom-right diagonal and coefficients of 0 everywhere else. The system would then appear like the following, in which the underlined blank denotes a number to be determined by our calculations, Chapter 3 Systems of Linear Equations and Inequalities 56 1X + 0Y + 0Z = __, 0X + 1Y + 0Z = __, 0X + 0Y + 1Z = __. With the system in diagonal form, we would be able to read the solutions directly as X = __, and so on. This is precisely how a computer or graphing calculator solves a system of linear equations using the elimination method. In the example that follows, we will use the row operations of the elimination method only as far as is necessary to eliminate the coefficient of X in equations B and C and the coefficient of Y in equation C. This will result in an equivalent system of equations that looks like 1X + 4Y + 2Z = 29, 0X + 1Y + __Z = ___, 0X + 0Y + 1Z = ___. A system of equations like this is said to be in triangular form or row-echelon form because the coefficients of the system form a diagonal of 1’s from upper left to lower right, with 0’s in the “triangle” below the diagonal. Notice that, when the system is in triangular form, the bottom equation says Z = ___. From that point on, we will use the technique of backward substitution, using the known value of Z in the second equation to solve for Y. With values for both Z and Y in hand, we solve for X in the first equation. To summarize, our method will be to use row operations to reduce the system of three equations to triangular form, having zeros below the main diagonal. Then we will start to read the solution from bottom to top, substituting known values as necessary. An example will show how the method works. Example 2 Elimination Method on Three Equations in Three Variables Solve the following system by the elimination method:[ED: Tabs removed; not needed for comp]] A. X + 4Y + 2Z = 29, B. X + 3Y + 3Z = 35, C. 3X + Y + 2Z = 33. Solution We begin by eliminating the coefficient of X in equation B. Equation B has a “1” coefficient on X, so the necessary row operation is to multiply equation A by <->1 and add the result to equation B: [comp: align each set of eq. on equal signs] <->1A <->1X – 4Y –2Z = <->29 +B 1X + 3Y + 3Z = 35 = B1 0X – 1Y + 1Z = 6 Likewise, we must eliminate the coefficient of X in equation C. Multiply equation A by <>3 and add that to equation C, getting the result, <->3A <->3X – 12Y – 6Z = <->87 +C 3X + 1Y + 2Z = 33 = C1 0X – 11Y – 4Z = <->54 Chapter 3 Systems of Linear Equations and Inequalities 57 To get a good look at our progress in the elimination method, we now rewrite the system of equations using B1 and C1 in place of the original B and C: A. 1X + 4Y + 2Z = 29, B1. 0X – 1Y + 1Z = 6, C1. 0X – 11Y –4Z = <->54. Clearly, the two row operations we just performed have taken us most of the way toward the triangular form. The next step is to eliminate the coefficient of Y (<->11) in the third row. A row operation using equation A applied to equation C1 would create a nonzero coefficient on the X variable in C1, which we want to avoid. However, row B1 has a 0 coefficient on X, so multiplying that equation by any number will result in a 0X being added to the 0X in equation C1, retaining the 0 coefficient that the triangular form calls for. Therefore, we use equation B1 for the row operation of the next step. Multiply B1 by –11, and add it to C1 as follows: <->11*B1 0X + 11Y –11Z + C1 0X – 11Y – 4Z = C2 0X +0Y <-> 15Z = <->120 = <->66 = <-> 54 To complete our work on C2, we divide both sides of the equation by <->15, C3. 0X + 0Y + 1Z = 8. Now, turn to equation B1. We want the coefficients of X and Y to be 0 and 1, respectively. Multiplying by –1 will give us an equation B2 with the desired coefficients: B2. 0X + 1Y – 1Z = <->6. Collecting the new equations B2 and C3 with equation A, we have the system, A. 1X + 4Y + 2Z = 29, B. 0X + 1Y – 1Z = <->6, C3. 0X + 0Y + 1Z = 8, X+ or 4Y + 2Z Y–Z = <->6; Z = 8. This system is in triangular form. Use back-substitution to solve the system. Substituting Z = 8 into equation B2, we solve for Y: Y – 8 = -6, Y = 2. Substituting Y = 2 and Z = 8 into equation A, we solve for X: X + 4(2) + 2(8) = 29, X + 24 = 29, X = 5. CHECK THE SOLUTION A. 1X + 4Y + 2Z = 29 5 + 4(2) + 2(8) = 29 5 + 8 + 16 = 29 ✓ B. 1X + 3Y + 3Z = 35 5 + 3(2) + 3(8) = 35 5 + 6 + 24 = 35 ✓ C. 3X + 1Y + 2Z = 33 = 29; Chapter 3 Systems of Linear Equations and Inequalities 58 3(5) + 2 + 2(8) = 33 15 + 2 + 16 = 33 ✓ Make a Note For solving a system of three or more equations with the same number of variables, the elimination method is used to transform the system of equations into triangular form. That reveals the value of one variable, and then the other variables can be calculated by backward substitution. Elimination Method in Matrix Form The elimination method consists of elementary row operations (multiplying an equation by a nonzero number, adding one equation to another) applied to the coefficients and right<->hand side of a system of equations. The names given to the variables, such as X, Y, or Z do not matter. In fact, we will now see that it is easier to work the elimination method when the system of equations is written as an array of coefficients and right-hand-side values without the variable names. A matrix is a rectangular array of numbers. A system of equations can be written in matrix form by arraying the coefficients and right-hand-side numbers in their natural place. For example, the system of equations we used in Example 2 would be expressed in matrix form, with the rows (equations) labeled R1, R2, and R3 and the right-hand-side numbers separated from the coefficients by vertical lines, as follows: 1 X + 4Y + 2Z = 29 R1 1 4 2 29 1 X + 3Y + 3Z = 35 R2 1 3 3 35 3 X + 1Y + 2 Z = 33 R3 3 1 2 33 Elementary row operations are easy to perform on the system in matrix form. For example, to indicate that you are doing the row operation “multiply row 1 by –1 and add to row 2,” you could simply make the notation “–1R1 + R2” and give the resulting system of equations in matrix form: –1R1+R2 1 4 2 29 0 −1 1 6 3 1 33 2 The next row operation in the solution of that example was to multiply row 1 by –3 and add the result to row 3:[comp: for matrices throughout, align each column of numbers within the matrix.] <->3R1 + R3 1 4 2 | 29 0 <->1 1 | 6 0 <->11 <->4 | <->54 The third step was to multiply row 2 by –11 and add to row 3: <->11R2 + R3 1 4 2 | 29 0 <->1 1 | 6 0 0 <->15 | 120 Chapter 3 Systems of Linear Equations and Inequalities The final step to triangular form was to multiply row 3 by 1 4 2 <->1R2 0 1 <->1 | <->6 −1 R3 15 0 0 1 | 8 59 −1 and row 2 by –1: 15 | 29 Here, remembering that the columns of the matrix are identified with the variables X, Y, Z, and the right-hand side, we read part of the solution from the bottom row, Z = 8, and recover the other solutions by substitution. Make a Note Row Operations on a System in Matrix Form Each row in the matrix form corresponds to one equation. The elementary row operations are the three transformations that do not change the solution to the system: 1. Multiplying an entire row by a nonzero number. 2. Adding one row to another. 3. Interchanging the order of the rows in the matrix. When the elimination method is programmed into a computer or a graphing calculator, the method continues beyond the triangular form, eventually reducing the matrix so that it has 1’s on the diagonal coefficients (upper-left to lower-right), 0’s for all other coefficients, and numbers in the right-hand-side values. The matrix form of our example would eventually look like 1 0 0 5 0 1 0 2. 0 0 1 8 In that reduced row-echelon form (rref), you can read the solution directly. The first three columns are for X, Y, and Z, so the solution reads X = 5, Y = 2, and Z = 8. One of the extraordinary abilities of a graphing calculator is to solve a system of equations by transforming it through row operations to the rref routine. Learning with Technology Using a Graphing Calculator to Solve a System of Equations 1. Set up a blank matrix. After calling up the entry screen for the Matrix tool, choose the Edit option. You must first select a name for the matrix (we’ll call it "A"). The calculator will also ask you to define the number of rows (equations) and columns (variables + right-side) in the matrix. A matrix with 3 rows and 4 columns is called a “3-by-4” matrix, often written in shorthand as “3 4.” A 3 4 array may not show up entirely on one screen. In the screen shown here, the underline ("_") symbols in the far right column indicate that more of the array is off the right side of the screen. {INSERT FIG. UN3.2a} 2. Edit the system matrix. Enter the coefficients and right-hand-side numbers from the system of equations into the corresponding cells of the matrix. The example here shows columns 2, 3, and 4. The _ to the left of the first column indicates that some of the matrix is off the screen to the left. {INSERT FIG. UN3.2b} Chapter 3 Systems of Linear Equations and Inequalities 60 3. Invoke the rref routine. Expanded form moved above by 1st mention}. Return to the main screen, then call up the reduced row echelon form by invoking the calculator’s built-in routine for that procedure. It is accessed in the TI-83 model by the name rref as a Matrix/Math function on the Matrix entry screen. as shown here. {INSERT FIG. UN3.2c} After the parenthesis of the rref instruction, insert the matrix name A by entering it from the Matrix/Names screen: {INSERT FIG. UN3.2d} Hitting Enter at this point will invoke the calculator’s rref routine to operate on your matrix A, giving the matrix form of the solution to the system of equations. In the final screen, shown here, one has to remember how each column is interpreted: the first three columns are the coefficients of the variables X, Y, and Z, and the rightmost column is the solution column. Thus, we read the system as saying 1X = 5, 1Y = 2, and 1Z = 8. {INESRT FIG. UN3.2e} We are now in a position to solve the problem that was posed in the Spotlight for this section. Recall that Carol’s task was to determine how many inquiries to Metropolitan University’s graduate certificate program in management were generated from the advertisements in each of four different newspapers and magazines. Example 3 Metropolitan University’s Advertising When we left Carol in the Spotlight, she was about to show the assistant director of Admissions how she could find out how effective his various advertisements had been over the past four months. Mr. Martinez gave Carol the following information about the history of the advertisements and the inquiries received. Source: Fortune Chicago Chicago Magazine Crain’s Chicago Business Suburban papers Price/ad: $3,500 $3,700 $2,300 $2,400 Total monthly cost Inquiries received August ads 1 0 1 3 $13,000 520 Septembe r. ads 0 1 1 2 $10,800 445 October ads 0 1 0 2 $ 8,500 300 Novembe r ads 1 0 1 2 $10,600 440 Carol had in mind a linear model for the effects of any one magazine’s advertising. This means that she was assuming that any one source (magazine or newspaper) would always produce the same number of inquiries per advertisement, from one month to the next. The linear model also implies that if an ad appeared 3 times in the suburban papers in August, it should produce 1.5 times the number of inquiries as when only 2 ads were placed, which was the case in September, October, and November. Chapter 3 Systems of Linear Equations and Inequalities 61 We shall use the letters r1 to r4 to represent the response rate for each of these four publications. The response rate is the number of inquiries per ad, per month, for that publication. Carol’s linear model of the effects of advertising on inquiries can be written in the following equation: r1(Fortune ads) + r2(Chicago Magazine ads) + r3(Crain’s ads) + r4(suburban ads) = total inquiries. Under Carol’s assumption of a linear model, this equation will be the same for every month. In the month of August, for example, Metropolitan University took out 1 Fortune ad, 1 Crain’s ad, and 3 suburban ads, and the result was 520 inquiries. Therefore, Carol’s linear model supposes that the number of inquiries will satisfy the equation, r11 + r20 + r31 + r43 = 520 (August inquiries). Similarly, her model supposes that r10 + r21 + r31 + r42 = 445 (September inquiries), r10 + r21 + r30 + r42 = 300 (October inquiries), r11 + r20 + r31 + r42 = 440 (November inquiries). Think about This Carol has made a mathematical model of the process by which advertisements result in inquiries to the Admissions Office. This model is only as good (accurate) as the assumptions that go into it. Take a moment to think about how reasonable her assumptions are. Consider the following questions that challenge the linear model. 1. Can the effect of an advertisement be different from month to month? Why or why not? 2. Can the effect of more ads in a month in the same magazine be more or less than would be predicted by multiplying the assumed “response rate per ad” by the number of ads? Why might the effect be more? Why might it be less? 3. Can the effect of advertising in one magazine make the advertising in another magazine, in the same month, more or less effective than otherwise? What would make it more effective? What might make it less effective? As you probably can see, each of Carol’s assumptions is open to some challenge. But each of the assumptions is still reasonable enough for her to proceed with the linear model and see if the results make sense, especially since Mr. Martinez doesn’t have any alternative right now. Given Carol’s model, solve for each of the response rates to determine the effectiveness of advertising in each publication. Solution Putting the coefficients and totals in matrix form, with a top row reminding us of the variables, we have r1 r2 r3 r4 A. 1 0 1 3 520 B. 0 1 1 2 445 C. 0 1 0 2 300 D. 1 0 1 2 440 Chapter 3 Systems of Linear Equations and Inequalities 62 The first step toward getting the matrix into triangular form will be to eliminate the 1 under r1 in the bottom row. We do so by multiplying row A by <->1 and adding the result to row D. This gives us a new matrix, r1 r2 r3 r4 A. 1 0 1 3 520 B. 0 1 1 2 445 C. 0 1 0 2 300 D1. 0 0 0 <->1 <->80 That was real progress because we can already see the equation <->r4 = <->80 emerge in row D1. This will tell us that each advertisement in the suburban papers produces 80 responses. Continuing with the elimination method, we must eliminate the 1 under r2 on row C. Doing so by the appropriate row operation (<->1 row B added to row C) gives us a new table: r1 r2 r3 r4 A. 1 0 1 3 520 B. 0 1 1 2 445 C1. 0 0 <->1 0 <->145 D1. 0 0 0 <->1 <->80 After multiplying rows C1 and D1 by <->1, the matrix will be in triangular form, so we can start the process of backward substitution to get the four response rates. Row D1 says <->1 r4 = <->80, so r4 = 80 responses per ad in the suburban papers. Likewise, row C1 reveals that r3 = 145 responses per ad in Crain’s Chicago Business. Using row B and our known values of r3 and r4 we determine the value of r2 as follows B. r2 + 145 + 2(80) = 445, so r2 = 140 (responses per ad in Chicago Magazine). Finally, we use row A to calculate r1, r1 + 0r2 + 1r3 + 3r4 = 520, r1 + 0 + 145 + 3(80) = 520, so r1 = 135 (responses per ad in Fortune’s Chicago edition). Learning with Technology 1. To use a graphing calculator to solve this system, what matrix dimensions would you enter? 2. Verify our solution using your graphing calculator. Answer: 1. 4 5 The conclusion of this linear model is that the number of responses per ad in Fortune, Chicago Magazine, Crain’s, and the suburban papers are 135, 140, 145, and 80, respectively. The first three response rates are quite similar, but the number of responses per ad in the suburban papers is much less than the others. This analysis has a very practical value. The original table of values given by Mr. Martinez shows that Fortune and Chicago Magazine ads cost about the same, but the Crain’s ad costs much less. When ads in those three magazines produce similar effects in terms of Chapter 3 Systems of Linear Equations and Inequalities 63 response rate, what does that tell us about where advertising is most effective per dollar spent? And look at the suburban papers. How does their effect per dollar spent compare with the other sources? INCONSISTENT AND DEPENDENT SYSTEMS Before leaving the topic of systems of linear equations, let's look at two other possible outcomes of the elimination method. When we solve a system that has the same number of equations as the number of unknowns, we expect to get a solution where there is one value for each variable. But in section 3.1 of this chapter we found that there are two situations where this kind of solution will not occur. As you will recall, if we visualize a system with two variables and two equations as two lines on the coordinate plane, the system has no solution if the two lines are parallel and therefore do not intersect. In this case, the equations are inconsistent with each other. Another possibility is that the two lines will actually be one and the same, in which case all points on that line are solutions to the system of “two” equations. That is the situation of a dependent system of equations. With more than two variables, it is not as easy to visualize the system of equations. A linear equation in three variables can be pictured as a plane in three-dimensional space. Three such equations can be drawn as three planes, and if they are all tilted in different directions they will intersect in a single point in three-dimensional space, which is the solution to the system of three equations in three variables. But there are various situations where there is no solution, or no common point of intersection. Can the elimination method reveal when a system of equations has no solution or has many solutions? The answer is “yes” because the effect of the row operations that would otherwise produce a triangular array of coefficients in a table will produce a different result instead. Let’s look at two examples. Example 4 Inconsistent System An inconsistent system has no solution. As an example, attempt to solve the following system of equations using the elimination method: A. 2X + 3Y + Z = 25, B. 4X + Y + 5Z = 40, C. 2X – 2Y + 4Z = 10. Solution To work the elimination method most rapidly, we put this system in matrix form, where the columns correspond to the variables X, Y, and Z and the right-hand sides of the equations: A. B. C. 2 4 2 3 1 <->2 1 5 4 25 40 10 In the first sequence of steps, we will multiply row A by 1/2 so that X has a coefficient of 1 on the diagonal. Then we will multiply the new first row A1 by <->4 and add it to row B. Next, we multiply A1 by <->2 to add to row C to eliminate the X coefficients in those two rows. The result of those three operations is the following: A1. B1. C1. 1 0 0 1.5 <->5 <->5 0.5 3 3 12.5 <->10 <->15 In the matrix above, we can see trouble emerging. Rows B1 and C1 have identical coefficients but different right-hand-side values. It will be impossible to find values of Y and Z Chapter 3 Systems of Linear Equations and Inequalities 64 that satisfy both equations simultaneously. As we proceed with the elimination method, this will become even more apparent. Continuing, we divide row B1 by <->5 to produce a 1 coefficient for Y on the diagonal B2. Then we multiply B2 by 5 and add the result to row C1 to get the following table: A1. B2. C2. 1 0 0 1.5 1 0 0.5 12.5 <->0.6 2 0 <->5 The last equation says that 0X + 0Y + 0Z = <->5, which is impossible for any real numbers X, Y, and Z to satisfy. Make a Note The appearance of a row of zeros in a table of coefficients, together with a nonzero right-hand-side value, is the signal that the system of equations has no solution. Example 5 Dependent System A dependent system has more than one solution. This is indicated when two rows become identical, which means they have the same coefficients and the same right-hand-side value. At the end of the elimination method, this results in a row of all zeros. Consider this example: A. 2X + 3Y + Z = 25 B. 4X + Y + 5Z = 40 C. 2X – 2Y + 4Z = 15 We put this system in matrix form. A. B. C. X 2 4 2 Y 3 1 <->2 Z 1 5 4 25 40 15 Multiply row A by 1/2, then add –4 times the new row A1 to row B, and <->2 A1 to row C in order to eliminate the X coefficients on rows B and C The result of those three operations follows: A1. B1. C1. X 1 0 0 Y 1.5 <->5 <->5 Z 0.5 3 3 12.5 <->10 <->10 Divide row B1 by <->5 to produce a Y coefficient of 1 on the diagonal B2. Then multiply that row by +5, and add the result to row C1 in an attempt to eliminate the coefficient of Y (C2). We get the following matrix: A1. B2. C2. X 1 0 0 Y 1.5 1 0 Z 0.5 <->0.6 0 12.5 2 0 The last equation says that 0X + 0Y + 0Z = 0, which can be satisfied by many different values of X, Y, and Z. Chapter 3 Systems of Linear Equations and Inequalities 65 Make a Note The appearance of a row of all zeros in the table, including a zero for the right-hand-side value, is the signal that the system of equations has more than one solution. In this chapter, we will not consider how to treat these two special cases any further. However, in a course on matrix algebra (linear algebra) you would learn more about these, including how to “solve” a system of equations that has no solution and how to describe all the solutions to a system that has many solutions. SUMMARY This section of the text introduced a topic that has extremely broad usefulness across many fields of study and professions. The application exercises in this section will demonstrate the range of uses for systems of linear equations. The topic is easy to describe: you have several different relationships among the same set of variables. Each relationship corresponds to an equation, so the mathematical description of the situation is a system of several equations using several variables. In this section, we looked only at the problem where the number of variables is the same as the number of equations and all the equations are linear. This problem almost always has exactly one solution, which you can calculate using the elimination method. Each step of the elimination method is an elementary row operation, which can be the multiplication of a row by a nonzero number, the addition of one row to another, or an interchange of the order of rows in the matrix. Elementary row operations do not change the solution to the system, yet they simplify the system of equations. The end result of the elimination method is a system of equations in reduced row-echelon form, where the only nonzero coefficients are 1’s along the main diagonal of the matrix (from the upper-left, down and to the right). In that form, it is possible to read the solution of the system directly from the right-most column of the matrix. EXERCISES Skill Builders Problems 1–5 should be solved in sequence using the following system of equations. A. <->X B. 3X C. 2X + Y + 2Z =3 – 2Y + 4Z = 19 + 3Y – 5Z = <->19 1. What operation on row A would enable you to eliminate the coefficient of X in row B? Perform that operation to obtain a new row B, and label it B2. 2. What operation on row A would enable you to eliminate the coefficient of X in row C? Perform that operation to obtain a new row C, and label it C2. 3. What operation on row B2 would enable you to eliminate the coefficient of Y in row C2? Perform that operation to obtain a new bottom row, and label it C3. 4. What value of Z is implied by equation C3? Substitute that value of Z in equation B2, and solve for Y. 5. Substitute the values of Z and Y found in problem 4 into equation A and solve for the remaining unknown, X. Problems 6 – 10 should be solved in sequence using the following system of equations. Chapter 3 Systems of Linear Equations and Inequalities 66 A. X + Y B. <->X C. 2X + 2Z +2Y + 3Y =2 + 4Z = <->2 – 4Z = <->1 6. What operation on row A would enable you to eliminate the coefficient of X in row B? Perform that operation to obtain a new row B, and label it B2. 7. What operation on row A would enable you to eliminate the coefficient of X in row C? Perform that operation to obtain a new row C, and label it C2. 8. What operation on row B2 would enable you to eliminate the coefficient of Y in row C2? Perform that operation to obtain a new bottom row, and label it C3. 9. What value of Z is implied by equation C3? Substitute that value of Z in equation B2 and solve for Y. 10. Substitute the values of Z and Y found in problem 4 into equation A, and solve for the remaining unknown, X. Use the elimination method to solve the following systems of equations. Check your answer by substituting values back into the original equations. 11. A. 1X + 0Y + 1Z = 5 B. 0X + 1Y – 2Z = <->10 C. 2X + 5Y + 3Z= 4 12. A. 1A + 2B + 1C = 7 B. –1A – 2B + 4C C. 0A + 2B + 3C= 8 =3 13. A. 1A + 2B – 1C = 7 B. –1A + 5B + 2C = 15 C. 2A – 6B + 1C= <->13 14. A. <->3X + 3Y – 5Z = <->6 B. 2X – 1Y + 2Z = <->1 C. –1X + 0Y + 3Z = 13 15. A. B. C. 16. A. B. C. 1A + 3B + 1C = 7 0A + 5B + 2C = 11 2A + 4B – 3C= <->3 1X + 0Y + 1Z = 3 1X – 1Y + 1Z= 0 3X – 2Y + 5Z= 11 17. A. 2A + 3B + 4C = 0 B. 5A + 6B + 8C = 0 C. 4A – 2B + 3C= 0 18. A. 1A + 1B + 1C= 3 B. 2A + 3B – 1C= – 2 C. 3A + 0B + 1C= 0 19. A. 1X+ 2Y + 1Z = 7 B. 1X – 1Y + 1Z = – 2 C. –1X + 1Y – 2Z =6 20. A. 1A + 1B – 1C= 0 B. 0A + 1B + 1C = 0 C. 1A + 2B + 2C= 6 For each of the following determine whether the system is inconsistent (has no solution) or dependent (has more than one solution). Chapter 3 Systems of Linear Equations and Inequalities 67 2 X + 3Y + 4Z = 1 21. 4 X + 6Y + 8Z = −1 4 X − 2Y + 3Z = −3 1X + 2Y − 4Z = 11 22. − 2 X + 3Y − 6Z = 6 3 X + 4Y − 8Z = 13 1X + 1Y − 1Z = 0 23. 0 X + 1Y + 1Z = 0 1X + 2Y + 0Z = 0 Applications and Analysis When you solve a system of equations in the following section, use the elimination method. Show clearly which equation or table row you are multiplying by what number and adding to which other equation as you proceed through the steps of the elimination method. 24. Metropolitan University (part B). Pretend that it is the first week of December. Carol Haines is in the Admissions Office going through the mail, and Mr. Martinez walks in. “Carol, am I glad to see you. Say, I noticed on my November financial report that we had spent $13,000 on advertising rather than the $10,600 that I told you earlier in the week. I checked, and it turned out that we had run 3 ads in the suburban papers, not 2. And on top of that, my secretary just brought in a batch of inquiries that had come in the mail while she was ill last week. Her temporary replacement didn’t know what to do with them and just set them aside. There are 40 inquiries in that batch, which brings the November total to 480, not 440. How much of a difference do you think these changes will make in your estimates of the effectiveness of advertising in each of our sources?” “That depends,” she replies. “I didn’t run a sensitivity analysis on the original model. That is, I never tried to change the numbers slightly to see what effect such changes would have. But the problem was rather quick to solve. It will take me only a minute to change those two numbers and give you the results. Do you want to see what it looks like?” “Please,” said Mr. Martinez. Carol changed the November inquiries from 440 to 480 and the number of suburban ads in November from 3 to 2. This gave her the table shown here. r1 r2 r3 r4 1 0 1 3 520 0 1 1 2 445 0 1 0 2 300 1 0 1 3 480 Try to solve this system of equations using the elimination method. What has happened? 25. Restaurant staffing. Many restaurants are open long hours during the day, but only in certain times of the day is the restaurant extremely busy. As a consequence, a restaurant owner will typically hire waiters on several shifts, and the shifts are designed to overlap during the busy periods. In that way, the restaurant has many waiters on duty during the busy periods but fewer waiters on duty during the slow periods. Chapter 3 Systems of Linear Equations and Inequalities 68 Suppose that Fred’s Seafood Restaurant is open from 11 a.m. to 11 p.m., serving lunch and dinner. Fred Trawler, owner of the restaurant, has set up four shifts for his waiters. In the table here, those shifts are denoted by a 1 in any time period that the shift’s staff are supposed to be on duty. A 0 indicates that the shift members are not on duty during that time period. A zero separating two 1’s down a column indicates a “split shift,” in which the waiter is given time off but is asked to come back to the restaurant later in the day. Fred has made the split shifts shorter in total hours than regular shifts due to the inconvenience of having to go off duty and come back on duty. Shift 4 is for part-time waiters, who are usually students at a nearby university. The rightmost column of the table also shows how many waiters will be required during each three-hour block of time during the work day. Time block Waiters Shift 1 Shift 2 Shift 3 Shift 4 required 11-2 p.m. 1 1 0 0 9 2–5 p.m. 1 0 1 0 4 5–8 p.m. 1 1 1 1 14 8–11 p.m. 0 0 1 0 3 A. Write an equation that says, “The total number of waiters on duty between 2 and 5 p.m. is equal to 4.” (Hint: Write S1, S2, etc., for the number of waiters assigned to shift 1, shift 2, etc.) Which shifts are on duty during 2-5 p.m.? Use the variable names S1, S2, etc., to write an expression for the total number of waiters on duty during 2-5 p.m. [nice edits} B. Use the elimination method to find the numbers of waiters that Fred should hire on each of the four shifts. If he must have whole numbers of waiters (not fractions), how many should he hire? C. A new enterprise has located nearby, and Fred has noticed a much larger lunch crowd together with a slightly larger evening crowd. He has decided he now needs 12 waiters for the 11-2 time block, 4 for the 2-5 time block, 15 for the 5-8 time block, and 3 for the 8-11 time block. Use the elimination method to find the number of waiters that Fred should hire for each shift. 26. Advertising effects and walk-in traffic. Lois Waxman is the owner of Somebody Cares Gift Shop. She advertises on KABC radio and in the local newspaper. But she also gets a lot of walk-in traffic because her store has been around for a while and the large sign out front reminds people where she is. However, she is not sure how much of her sales are due to walk-ins and how much is due to the advertising she is paying for. Over 3 months, Lois collected the following information that might help her understand where her sales are coming from. January: 2 ads on KABC, 3 in the local paper, and sales of $10,000 February: 1 ad on KABC, none in the local paper, and sales of $6,000 March: 1 ad on KABC, 2 in the local paper, and sales of $7,500. A. Let s1 denote the dollar sales per ad in KABC, s2 the dollar sales per ad from the local paper, and s3 the unknown amount of sales from foot traffic. Explain why the equation, 2s1 + 3s2 + 1s3 = 10,000 could be used to account for the sales in January. B. Write equations to account for the sales in February and in March. C. Solve the system of three equations in three unknowns. How much of Lois’s sales are due to walk-in traffic? 27. Emergency room medical staff. A hospital’s emergency room is open 24 hours a day. At Lakeview Hospital, the medical staff work 12-hour shifts 3 days a week. The shifts are 6 a.m. to 6 p.m., noon to midnight, 6 p.m. to 6 a.m., and midnight to noon. Experience has shown that the need for emergency medical staff is greatest during the noon to 6 p.m. period, Chapter 3 Systems of Linear Equations and Inequalities 69 due to work-related accidents and traffic emergencies. In the table, a 1 in the shift columns indicates that the staff on that shift are working during a particular period. The rightmost column shows the need for staff during the period. Staff Shift 1 Shift 2 Shift 3 Shift 4 required 6 a.m.–noon 1 0 0 1 21 noon–6 p.m. 1 1 0 0 35 6 p.m.–midnight 0 1 1 0 32 midnight–6 a.m. 0 0 1 1 18 A. Write an equation that says, “The total number of medical staff on duty between 6 a.m. and noon is equal to 21.” (Hint: Write S1, S2, etc. for the number of staff assigned to shift 1, shift 2, etc. Which shifts are on duty during the period 6 a.m. to noon?) B. Use the elimination method to try to find the numbers of medical staff that Lakeview Hospital should hire on each of the four shifts. What has happened? C. One way to eliminate the dependency in this problem is to modify the coefficient matrix so that not all shifts work continuous 12-hour blocks. But we shouldn’t split shifts that are this long (6 hours on, 6 off, 6 on, 6 off would be a killer.) So shorten one of the shifts to a single 6hour segment. What then is the solution to the assignment problem? Can you explain its peculiar feature? Analysis Interchanging rows and columns in the elimination method. Look back to the Metropolitan University example. We gave letter names to the variables describing responses per advertisement in the Fortune (r1), Chicago Magazine (r2), Crain’s (r3), and suburban newspapers (r4). With that notation, the equations in the system looked like this: August: 1r1 + 0r2 + 1r3 + 3r4 = 520, September: 0r1 + 1r2 + 1r3 + 2r4 = 445, October: 0r1 + 1r2 + 0r3 + 2r4 = 300, November: 1r1 + 0r2 + 1r3 + 2r4 = 440. 28. If we interchange the rows of this system of equations, it will change the appearance of the table of numbers that we start with to work the elimination method. For example, the system might be rearranged like this. September: 0r1 + 1r2 + 1r3 + 2r4 = 445, August: 1r1 + 0r2 + 1r3 + 3r4 = 520, November: 1r1 + 0r2 + 1r3 + 2r4 = 440, October: 0r1 + 1r2 + 0r3 + 2r4 = 300. Does rearranging the rows change the solution? If so, why? If not, why not? (Hint: To answer this question, don’t try to solve this system. Think about what it means to rearrange the rows and about the meaning of a solution to the system of equations.) 29. We could also have chosen a different order to write the variables in. For example, instead of the order Fortune, Chicago, Crain’s, suburban, we might have written it in the order Crain’s, suburban, Fortune, Chicago. In this order, the equations would appear as follows: September: 1r3 + 2r4 + 0r1 + 1r2 = 445, August: 1r3 + 3r4 + 1r1 + 0r2 = 520, November: 1r3 + 2r4 + 1r1 + 0r2 = 440, Chapter 3 Systems of Linear Equations and Inequalities 70 October: 0r3 + 2r4 + 0r1 + 1r2 = 300. Rearranging the order of the variables in a system of equations has same effect as interchanging the sequence of columns of the table that we use for the elimination method. Does rearranging the columns of the table in this way change the solution? If so, why? If not, why not? (Hint: Again, don’t try to solve this system. Think about what it means to rearrange the variables, and what is the meaning of a solution to the system of equations.) 30. Rearranging rows and columns can put this table into upper-triangular form simply and without any calculations. In this example, there may be numbers other than 1 along the diagonal, but the table will have a triangular form. A B C D W 0 1 0 6 X 2 4 7 3 Y 5 0 0 <->2 Z 0 2 0 0 100 180 140 90 A. Find a way to interchange rows and columns so that a triangular form appears. (This is like working a puzzle. Hint: how many zeros should there be at the bottom row of the table in triangular form, and in which columns should they be placed? What about the next-to-bottom row?) B. When you have the table in triangular form, complete the solution of the system by backward substitution. Not all systems of linear equations can be prepared for quick solution in this way, but if you see a lot of zeros in a system table, it is worth taking a minute to try a few row or column switches. Graphing Calculator Exploration 31. Real estate prices. Daniel Marlin is a real estate agent in the seashore community of Bayside. His observation of the real estate market suggested to him that three features determine property prices: the floor area of the house, the amount of shoreline, and the acreage of the property. Daniel collected the following data on three houses that sold recently in Bayside. House Watson Carlisle Becker Floor (sq. ft.) 1,000 800 1,500 Shoreline (ft.) 75 50 100 Land (acre) 1 0.75 3 Selling price ($) 110,000 85,250 185,000 Use a graphing calculator to solve this system in matrix form. Write a linear equation that can be used to predict the selling price of any house in Bayside. Spreadsheet Exercise 32. The Microsoft ExcelTM spreadsheet program can solve systems of simultaneous linear equations. Excel is installed on so many computers and is used for so many purposes in business that it is a good program to know. Excel has a Solver routine to solve systems of linear equations as well as many other types of mathematical problem. The Solver is found under the Tools menu. Solve Exercise 18 above using the Solver, as follows. A B C D E F 1 0 0 0 Chapter 3 2 House 3 Watson 4 Carlisle 5 Becker Systems of Linear Equations and Inequalities 71 Floor (area) Shoreline Land Selling (ft.) (acre) price ($) 1,000 75 1 110,000 800 50 0.75 85,250 1,500 100 3 185,000 A. Set up the system of linear equations in matrix form beginning in row 2. Then add a row of zeros above Floor, Shoreline, and Land in cells B1 to D1, as shown. These cells will act as the unknown “prices” for Floor, Shoreline, and Land. B. Create a formula for cell F3 that would equal the price of the Watson property if the prices of floor, shoreline, and land are the values in cells B1 to D1. That is, for cell F3, enter = B$1*B3+C$1*C3+D$1*D3. The dollar signs in B$1 to D$1 will instruct Excel to hold row 1 when this formula is copied to other cells in the next step. A zero will appear in cell F3 because the current values of B1 to D1 are all zero. C. Copy the formula from cell F3 to cells F4 and F5. Zeros should appear in those cells, too. D. Call up the Solver from the Tools menu. The Solver Parameters box will appear. Clear the “Set Target Cell” and “Value of:” boxes. In the “By Changing Cells” box, type B1:D1. To the right of the “Subject to the Constraints” box is an Add button. Click on this. An “Add Constraint” box will appear. For “cell reference” give F3. Choose = from the menu. And for “constraint” give E3. This tells the solver to set the price of the Watson house (F3) equal to the number 110,000 that is in cell E3. Click “OK,” and the Solver parameters box will appear with one constraint in it, expressed as $F$3 = $E$3. Click the Add button again and use the same procedure to define the constraint F4 = E4. Click “OK.” Click the Add button a third time and define the constraint F5 = E5. Click “OK.” E. Now, in the upper right of the Solver Parameters box, click “Solve.” The solutions to this system of linear equations will appear in place of the zeros in cells B2 to D2. Save the spreadsheet in its present form. You can use it to solve any other 3 3 linear system simply by changing the numbers and labels in the table. CASE STUDY — APPRAISING REAL ESTATE A property owner who wants to sell a house often asks a real estate professional to determine a fair market price for, or appraise, the property. The usual method of a real estate appraiser is to find several houses that have sold recently in the area and that are comparable to the subject property on such features as the number of bedrooms and bathrooms, size of the house, existence of a garage, size of the lot, and so on. By looking at the selling prices of those comparables, the appraiser can come pretty close to estimating a fair market value for the subject property. {INSERT Photo 3.4} However, it is difficult, if not impossible, for the appraiser to find several houses that are exactly like the subject property and that have sold recently in the area. The appraiser must therefore make adjustments for slight differences between each comparable and the subject property. A comparable may have one more bedroom and a garage, another comparable may have one less bathroom, and so on. The appraiser has some sense of how much an extra bedroom, bathroom, or garage should be worth, and they use their sense of the “prices” for those features to make the necessary adjustment. As it is commonly practiced, real estate appraisal is rather judgmental and subjective. However, a mathematical analysis of sold properties can reveal the “prices” of the various features that the real estate appraiser is trying to take into consideration. In this case study, we’ll look at data on houses in Greenfield, Ohio, that were about to be sold and use that to Chapter 3 Systems of Linear Equations and Inequalities 72 estimate a selling price for one particular subject property. The data for this case study were taken from the web site of the Multiple Listing Service of Greater Cincinnati <http://www.cincymls.com/>. The subject property will be 11892 Maple Trail in Greenfield, Ohio. It has 3 bedrooms, 3 other rooms, 1 bathroom, a 1-car garage, and sits on 0.4 acres of land. Five comparables, on which other buyers had made offers (presumably at the asking price) are listed in the table. Other Garage Land Location Bedroom rooms Baths (cars) (acres) Price s ($) 12386 Cameron Dr. 2 2 1 1 0.2 56,000 330 4th St. 2 3 1.5 0 0.6 56,000 545 McKell Ave. 2 3 1.5 1 0.3 62,900 11395 Rowe Rd. 3 5 2 0 0.7 79,900 836 Jefferson St. 3 5 1 2 0.2 95,900 A. Your first task is to create a linear equation model for the price of a property in relation to the features of the property. Problem 14 of section 3.2 and problem 18 in this section use the same principle. The price of a property is written as a linear equation with five variables, which are the five features of the house given in the table above. The coefficient of each variable in the equation can be interpreted as the “price” of that feature. Using the data in the table, you can set up a system of five equations with the five unknown “prices.” Solving that system, you find the price for each feature, which gives you an equation for the price of any property. B. Do the “prices” you get for each feature seem reasonable? In particular, what is the price per acre for land in Greenfield, Ohio? What is the “price” for a garage (per car)? How much extra would a house cost if it had an extra bathroom? C. Use the equation from part A to predict the selling price of the subject property at 11892 Maple Trail. The owners listed it at (were willing to sell it for) $65,000. Do you think that was too much, too little, or just about right? D. The predictive accuracy of your mathematical model of the property price is limited by the range of features that it considers. What other features of a property might affect the selling price? Chapter Review The knowledge in this chapter is very practical. We began with the search for two unknowns that are related by two equations. The range of applications that can be written as a system of two equations in two unknowns includes demand and supply in economics, production constraints in management, and problems of parameter estimation and “curve fitting” in business, biology, psychology, and medicine. We used the word, "system" to describe the equations that we were solving. This word is used widely in mathematics. In Chapter 1, we encountered the “real number system.” In this chapter we have been studying “systems of equations.” The common meaning of the word "system" in all these instances is that of a collection or set that has some relationship among the parts. This meaning is important because when a problem can be written as a system of equations, the solution will be the values of several variables that solve each equation simultaneously. Graphically, each linear equation in a two-variable system corresponds to a line in the coordinate plane. So a pair of numbers that solves both equations correspond to a point in the coordinate plane that lies at the intersection of the two lines. Therefore, the graphical method to solve a 2 2 system of equations is to plot the lines on a coordinate plane and determine the coordinates of the point of their intersection. It is often faster, however, to solve the system algebraically. For 2 2 systems, we used the substitution method to find a solution in three steps. Chapter 3 Systems of Linear Equations and Inequalities 73 The solution to an inequality is quite different from the solution to an equation. Graphically, the solution of a linear equation is a line in the coordinate plane. The graphical solution to an inequality is a half of the coordinate plane that is bounded by (and may or may not include) the corresponding line. Another difference between equations and inequalities is that, whereas a system of equations in two variables can be solved only if there are exactly two equations, a system of inequalities can have a solution even if there are many inequalities to be satisfied. We noted in section 3.3 that a solution set may be a bounded region of the plane or an unbounded region, or there might be no solution at all. And we observed that some constraints in a system can be redundant, meaning that they can be added or removed without altering the solution. We also discussed the concept of slack in a constraint at a point and learned that slack tells us how much the right-hand side of a constraint can be changed before the constraint becomes binding. In the final section of this chapter, we used the elimination method to solve systems of linear equations that have more than two variables. To solve systems most efficiently, we performed elementary row operations on matrices consisting of the constants in the equations. A linear system that has the same number of equations as the number of variables ordinarily has exactly one solution. However, we did observe that an inconsistent system has no solution, revealed when two rows with identical coefficients have different solutions. In addition, we noted the possibility of infinite solutions in a dependent system, where row operations result in a row of zeros. CUMULATIVE TEST FOR CHAPTERS 1-3 1. Stockbroker earnings. A stockbroker in New York City is paid a base wage of $2,000 per month plus 30% of the commissions paid by her clients. The commission is 2% of the dollar amount of the clients’ stocks sold. A. Write an equation for the stockbroker’s total monthly earnings in relation to the dollar amount of stocks sold. Express the equation in simplest terms. B. Graph the equation for the dollar amount of clients’ stock sold ranging from zero to $500,000. C. Determine algebraically how much volume (dollar amount of sales) a stockbroker would have to generate in a month in order to earn $3,500 in total compensation? Verify your answer by showing that point on your graph. 2. Stockbroker earnings. Another stockbroker made the comment over lunch that his firm pays him 40% of his commissions and charges clients 2% of the dollar amount of stocks sold. However, he says he has to hustle each month because his base salary is quite low. Last month, he says, he earned $4,500 and sold $375,000 worth of stock for his clients. What is his base salary? 3. Monthly mortgage. The monthly payment on a mortgage loan for a house depends on the interest rate of the loan and the price of the house. For a $100,000 house loan, the monthly payment at an interest rate of 5% per year would be $536.82. At a 6% annual interest rate, the payment would be $599.55 per month. A. Create a linear equation that shows the monthly mortgage payment in relation to the interest rate. B. Using the linear equation from part A, estimate the monthly payment that would be required if the interest rate were 5.625% per year. 4. Temperature. The equation for temperature in degrees Fahrenheit (F) in relation to degrees centigrade (C) is F = 32 + 1.8C. What is the equation for the temperature in degrees centigrade, given the Fahrenheit temperature? Chapter 3 Systems of Linear Equations and Inequalities 74 5. Materials budget. The cost of the materials used in manufacturing one table is $120 and for one bookcase it is $45. A. Write an equation that shows the number T of tables and number B of bookcases that can be manufactured within a materials budget of $8,100, then graph that equation. B. Determine algebraically how many tables could be manufactured within a $8,100 materials budget if the firm had already committed to manufacturing 100 bookcases, then show your answer on the graph from part A. 6. Break-even analysis. A firm has to pay $6,500 per month on utilities and rent, even if it is not producing anything. The cost of materials, energy and labor used to create its product is $30 per unit produced. A. Write an equation that shows the total expenditure E of the firm in relation to the quantity Q of the product it manufactures, then graph that equation for values of Q between 0 and 200. B. The firm obtains sales revenue by selling the product for $85 per unit. Draw the firm’s revenue line on your graph from part A. Determine both algebraically and graphically what quantity the firm must produce per month in order to break even. 7. Diet planning. Kalista is contemplating a diet composed entirely of raisins and green peas. The table describes the amount of several ingredients in these two foods and the recommended daily allowances. Content per Raisins (1 Green peas (1 Recommended piece cup, or 145 g) cup, 160 grams) daily allowance Calories 435 125 Maximum 2,000 Protein (grams) 5 8 Minimum 46 g. Calcium 71 38 Minimum 1,200 mg. Iron (mg.) 3.0 2.5 Minimum 15 mg. Vitamin C 5 16 Minimum 60 mg. If Kalista wants to eat a diet that gives her exactly the minimum requirements of iron and Vitamin C, how many cups of raisins and how many cups of peas should she eat? 8. Admissions screening. The policy of Valley College is to admit an applicant automatically if their average SAT score is at least 550 and their high school GPA is at least 2.75, and 100 times the GPA added to the average SAT score is at least 925. A. Write the conditions for automatic acceptance in the form of a system of linear inequalities. B. Sketch the “automatic acceptance” region in the SAT-GPA plane. C. Are any of the constraints redundant? D. Is the automatic acceptance region bounded or unbounded? E. If an applicant has a GPA of 3.05 and an SAT of 600, how much slack will they have in each constraint? Will they be automatically admitted? 9. Advertising effectiveness. Randy and Martin have obtained a franchise to sell AmritaTM aroma diffusers in the St. Louis region. In their first 3 months of marketing, they have been using several media. Some people have commented that they have seen the Amrita ads in a local newspaper. Radio advertising also seems to be somewhat effective. Advertisements in a citywide magazine also have been producing some results. Randy and Martin take orders over a toll-free number that they publicize in their ads, but they have not been asking callers how they heard about the aroma diffuser. Thus, it seems difficult to know which advertisements are more effective. This table shows the number of advertisements Chapter 3 Systems of Linear Equations and Inequalities 75 taken per month in each medium together with the monthly total of calls to the toll-free number. Help Randy and Martin determine the effectiveness of their advertising. Advertisements Radio Newspaper Magazine Total calls February 1 3 2 500 March 0 2 2 380 April 2 4 1 500