Brigham Young University - Idaho College of Physical Sciences and Engineering Department of Mechanical Engineering Class Prep Notes #5 Root Finding and Iterative Solutions Finding the roots of an equation is a common engineering application. Many strategies exist for doing this. We will focus on three techniques in this class. The three techniques are: Plotting Bisection Method Excel’s Goal Seek Plotting Creating a graph is a relatively quick and easy process to approximate the roots of an equation. It is also commonly used as a starting process to an iterative method. Once the graph is created, the scale may easily be refined to zoom the plot in on the area of interest. Bisection Method You may note when using a graph to locate roots that f(x) changed signs on opposite sides of the root. In general, if f(x) is real and continuous in the interval from f(xlow) to f(xhi) and f(xlow)*f(xhi) < 0 then there is at least one real root between xlow and xhi. The bisection method is one type of incremental search method in which the interval is always divided in half. If a function changes sign over an interval, the function value at the midpoint of the interval is evaluated. The location of the root is then determined to lie within the subinterval in which the sin change occurs. The process is repeated by refining or halving the estimates. A simple algorithm for the bisection calculation follows: 1. Choose initial guesses, xlow and xhi, such that the function changes sign over the interval. This can be checked by ensuring that f(xlow)*f(xhi)<0. 2. Estimate the root xroot by finding the midpoint of the interval: xroot xlow xhi 2 3. Determine which subinterval the root lies in and adjust the limits to create a new interval If f(xlow)*f(xroot)<0 the root lies in the lower subinterval Set xhi=xroot If f(xlow)*f(xroot)>0, the root lies in the upper subinterval Set xlow=xroot 4. Repeat steps 2 through 4 until desired termination criteria is reached. Termination Criteria A couple of methods are commonly used for deciding when to end the iterations and accept the estimate of the root. The simplest method involves comparing computing the absolute value of the last two root estimates. The iterations stop when the desired threshold is reached. A tolerance threshold of 0.001 is reasonable for many applications. The second method involves calculating an error estimate as percentage and stopping when the desired error threshold is reached. The error estimate can be calculated from the following formula: x new x old root new root 100% xroot A reasonable error estimate for many applications is 0.1%. Limitations of Bisection Method The Bisection method will not solve double root problems where the curve touches rather than crosses the axis. This is due to the fact that the bisection method relies on the sign change that occurs when the axis is crossed. See the double root plot example shown below. Using Goal Seek The third method involves a what-if analysis tool called Goal Seek. Goal Seek takes a result and determines possible input values that produce that result. Excel help on Goal Seek is included below for your convenience. Use Goal Seek to find out how to get a desired result If you know the result that you want from a formula, but you are not sure what input value the formula requires to get that result, you can use the Goal Seek feature. For example, suppose that you need to borrow some money. You know how much money you want, how long a period you want in which to pay off the loan, and how much you can afford to pay each month. You can use Goal Seek to determine what interest rate you must secure in order to meet your loan goal. NOTE Goal Seek works with only one variable input value. If you want to determine more than one input value, for example, the loan amount and the monthly payment amount for a loan, you should instead use the Solver add-in. For more information about the Solver add-in, see the section Prepare forecasts and advanced business models, and follow the links in the See Also section.