Section 8: Basic Trigonometric Functions Our goal in this section is to define and develop the six basic trigonometric functions. We start with two of the more important transcendental functions, namely the sine function and the cosine function. These are universally denoted by y = sin x and y = cos x respectively. In order to define sin x and cos x, we first need to agree on how we are going to measure angles. An angle, θ, consists of two rays emanating from a common point called the vertex. One ray is called the initial side, the other is the terminal side. Because we can choose where to put the origin in the Cartesian plane and we can decide where the positive x-axis goes, we will assume that the vertex of our angle is at the origin and that the initial side is along the positive x-axis. This is called putting the angle in standard position. Finally, since only the direction of the rays matters, we can chop off both rays at length one. Our angle theta is now determined by the two points at the end of these line segments of length 1, both of which lie on the unit circle whose equation is x2 + y 2 = 1. The point on the initial side is simply (1, 0). Here we will denote the point on the terminal side by P (θ). All we need to determine how to measure θ is to know how we generated the angle. That is we need to know how we go from (1, 0) to P (θ). If we generate θ by moving from (1, 0) counterclockwise (CCW), we say θ has positive measure. If we generate θ by moving from (1, 0) clockwise (CW), we say that θ has negative measure. The radian measure of θ is the length of the arc on the unit circle we travel from (1, 0) to P (θ) when generating θ. We use a minus sign in front when θ is generated CW. Since the unit circle has circumference 2π · 1 = 2π, once around the circle CCW is 2π radians. Halfway around the circle CCW is therefore π radians. One fourth the way around 2π π the circle CCW is = radians. So we see that P (θ) = P (ϕ) any time that θ and ϕ have 4 2 measures which differ by a multiple of 2π radians. So we write P (θ) = P (θ + 2kπ) for all k ∈ ZZ. Two such angles are called coterminal. We call the x-coordinate of P (θ) the cosine of θ and denote it by cos θ. The y-coordinate of P (θ) is the sine of θ and is denoted by sin θ. It is fairly clear that these are functions of θ where θ, as a length, can be any real number. Therefore both the sine function and cosine function have domain IR, unless restricted for some reason. Since all the points on the unit circle have coordinates bounded between 1 and −1, we have that −1 ≤ sin θ ≤ 1 and −1 ≤ cos θ ≤ 1. By the horizontal line test every one of these 29 Section 8: Basic Trigonometric Functions values is taken on, so the sine function and the cosine function both have range [−1, 1]. These serve as our first examples of functions with bounded range. Meaning there is a number M with f (x) ≤ M , for all x ∈ dom f . Since P (θ) = P (θ + 2kπ) for all k ∈ ZZ, we have sin(θ) = sin(θ + 2kπ), and cos(θ) = cos(θ + 2kπ) for all k ∈ ZZ. A function which repeats its values regularly is called periodic. The period of a periodic function is the length of the smallest continuous input interval between repetitions. Here we see that sin x and cos x have period 2π. Thus we would not be able to describe the behavior of sin x as x −→ ∞ except to say that the values oscillate. Neither the sine function nor the cosine function have any asymptote. The cosine has maximum value 1 and attains this value at θ = 0 + 2kπ, where k ∈ ZZ. The cosine has minimum value −1 and attains this at θ = π + 2lπ = (2l + 1)π, where l ∈ ZZ. The sine function attains maximum value of 1 when θ = π/2 + 2kπ = (4k + 1)π/2, where k ∈ ZZ. The minimum value of the sine function is −1 and occurs when θ = 3π/2 + 2lπ = (4l + 3)π/2, where l ∈ ZZ. For a given point P (θ) on the unit circle there are several other points which are natural to consider. The first is the antipodal point on the opposite side of the unit circle. This point has coordinates (− cos θ, − sin θ). It is also P (θ + π). We thus derive the relationship π sin(θ + π) = − sin θ. Another point to consider is radians CW from P (θ). It has coordinates 2 (sin θ, − cos θ) and is P (θ − π/2). Thus cos(θ − π/2) = sin θ, which means that the graph of the sine function is the graph of the cosine function shifted right π/2 units. Equivalently the graph of the cosine function is the graph of the sine function shifted left π/2 units and we have that cos θ = sin(θ + π/2). In short to determine the graphs of these functions, it suffices to graph the sine function for 0 ≤ θ ≤ π and apply all the relations we found above. This is made even easier when we use the fact that the unit circle is symmetric about the y-axis, giving us sin θ = sin(π − θ). So we only need to determine values for 0 ≤ θ ≤ π/2. Before we do so we remark that the x-intercepts of the sine function have x-coordinates nπ, where n ∈ ZZ. The x-coordinates of the intercepts of the cosine function are (2m + 1)π/2, where m ∈ ZZ. Now sin 0 = 0, and sin π/2 = 1. Also as θ increases from 0 to π/2, the value of the sine √ function will increase. In the next section it is derived that sin π/6 = 1/2, sin π/4 = 2/2, 30 Section 8: Basic Trigonometric Functions and sin π/3 = √ 3/2. With just these five values (which you should memorize), we can sketch a remarkably good graph of z = f (θ) = sin θ. We can also therefore graph its shift left by π/2 units, z = cos θ. The reciprocal of the sine function is the cosecant function, and is denoted csc x. It has a vertical asymptote when sin x = 0. It has a relative minimum when sin x = 1, and a relative maximum when sin x = −1. It is also periodic with period 2π. The reciprocal of the cosine function is the secant function, and is denoted sec x. It has period 2π, and is a shift left π/2 units of the cosecant function. The ratio of the sine function to the cosine function is the tangent function, denoted tan x = sin x/ cos x. This function is zero when sin x = 0, i.e. x = nπ, where n ∈ ZZ. It has a vertical asymptote whenever cos x = 0, which is when x = (2m + 1)π/2, where m ∈ ZZ. Since −/− is +, the tangent function has period π. As the ratio of an odd function and an even function the tangent function is odd. Finally as x −→ (π/2)− , sin x −→ 1, and cos x −→ 0+ , so tan x −→ ∞. On the other hand as x −→ (π/2)+ , tan x −→ −∞. The reciprocal of the tangent function is the cotangent function, denoted cot x. Collectively these make up the six basic trigonometric functions. We normally concentrate on sin x, cos x, tan x, and sec x. We will discuss inverse trigonometric functions, and the difference quotients of trigonometric functions in later sections. Several last points need to be made here. First of all, for any θ ∈ IR, there is a value r ∈ IR, called the reference angle, so that 0 ≤ r ≤ π/2, with | sin(θ)| = sin r and cos r = | cos(θ)|. For r 6= 0, π/2, there is therefore a unique right triangle of hypotenuse 1 which has base cos r and height sin r. This is the reference triangle for θ. We investigate this connection with right triangles further in a later section. Second, since the point P (θ) lies on the unit circle we have that (sin θ)2 + (cos θ)2 = 1, for all θ ∈ IR. Since this follows from the Pythagorean Theorem, this is called the Pythagorean Identity. This is one of the more important of hundreds of trigonometric identities. Exercise: Read sections 9 and 10. 31 Section 9: Special Values of Trigonometric Functions A crucial theorem in trigonometry is Euclid’s Theorem on similar triangles. The triangle with vertices at A, B and C is denoted by 4ABC. Two triangles 4ABC and 4DEF are similar provided that their interior angles have the same measure. So the angle with vertex at A has the same measure as the angle with vertex at D, etc. We state the theorem without proof. Theorem: (Euclid) 4ABC with side lengths a = |AB|, b = |BC|, and c = |AC| is similar to a b c 4DEF with side lengths d = |DE|, e = |EF | and f = |DF | if and only if = = . d e f This theorem allows us to define the trigonometric functions using P (θ). First we draw a line which touches the unit circle only at P (θ). Such a line is said to be tangent to the curve at the point P (θ). (You might be asking how this can be done.) Except when the line is horizontal or vertical (i.e. θ is a multiple of π/2) it has both an x-axis intercept D, and a y-intercept E. Call the origin O, and put B = (cos θ, 0). Then 4OP B is similar to 4ODP . |OD| |OP | |OD| 1 Therefore = . That is, = . So since sec θ and cos θ have the same sign |OP | |OB| 1 | cos θ| the coordinates of D are (sec θ, 0). Similarly the coordinates of E are (0, csc θ). A vertical line is tangent to the circle only when θ = kπ. For these values of θ the sine function is zero. Hence the cosecant is undefined. Notice though that the coordinates of D are still (sec θ, 0). The anagolous result holds true for horizontal tangents to the unit circle. |P D| |P B| |P D| | sin θ| To define the tangent we note that = . Which is to say = . So |OP | |OB| 1 | cos θ| the length from P to D measures the absolute value of the tangent of θ. In fact we see that the tangent line to the curve at P (θ) is perpendicular to the line, l, through O and P , so the slope of the tangent line is the negative reciprocal of the slope of l. But the slope of l is simply sin θ (rise over run) . So the value of tan θ tells us how to draw the tangent line to the unit cos θ circle at P (θ)! The point is that the definitions we made for trigonometric functions in section 8 did not fall out of a hat, they have geometric significance. Another application of Euclid’s Theorem is that given any right triangle with an included angle θ, where 0 < θ < π/2, we can define the six trigonometric functions as ratios of the lengths of the sides of the triangle. To do this draw the given triangle with the right angle in the lower right hand corner and θ in the lower left corner. Let the length of the base be a, the height be b, and the length of the hypotenuse be c. The other triangle is drawn with θ in 32 Section 9: Special Values of Trigonometric Functions standard position in IR2 with vertices (0, 0), (cos θ, 0), and (cos θ, sin θ). This triangle has base of length cos θ, height sin θ, and hypotenuse of length 1. By Euclid’s Theorem the ratios of corresponding side lengths are equal so we have 1 sin θ cos θ = = c b a So given two sides, for example a and c, we have an equation relating the two sides which can rewritten as an equation describing their ratio. In this case 1/c = cos θ/a implies a/c = cos θ. Equivalently c/a = 1/ cos θ = sec θ. The other formulas are b/c = sin θ, which is equivalent to c/b = csc θ, and b/a = tan θ, which is equivalent to a/b = cot θ. The famous mnemonic here is “soh-cah-toa” which is an acronym for “sine is opposite over hypotenuse–cosine is adjacent over hypotenuse– tangent is opposite over adjacent.” The sum of the measures of the angles of a triangle is 180◦ , which is π radians. So for a right triangle with included angle θ the other nonright angle ϕ must have measure (π/2) − θ radians. We call the two angles θ and ϕ complementary. Notice that the value of f (ϕ) = cof (θ) for any of the six trigonometric functions f , when defined as ratios of lengths of sides. Since the complement of the complement of θ is simply θ we get the mnemonic device “co-co=no-co”. The fact that a triangle has three sides explains why there are three pairs of complementary trigonometric functions. Finally, to demonstrate the utility of these concepts we derive some special values of the basic trigonometric functions. √ First take an isosceles right triangle with side s. The hypotenuse has length s2 + s2 = √ √ 2s2 = s 2. Also the included acute angles must be equal and have measure π/4. Therefore √ √ √ √ cos π/4 = s/s 2 = 1/ 2 = 2/2 = sin π/4. Also tan π/4 = 1 = cot π/4, and sec π/4 = 2 = csc π/4. Moreover any value of θ with reference angle π/4 has |f (θ)| = f (π/4), where f is any one of the six trigonometric functions. So to determine the actual value of f (θ) we simply need to know whether its positive or negative, which can be done since we know what quadrant the point P (θ) lies in, so we know the sign of its coordinates sin θ and cos θ. This is sufficient since all of the other basic trigonometric functions are defined in terms of sine and cosine. Second take an equilateral triangle of side s, drawn with horizontal base. The interior angles are all π/3. If we drop a perpendicular from the top vertex to the base, we split the triangle into two right triangles. The one on the left has base s/2 and hypotenuse s. By the 33 Section 9: Special Values of Trigonometric Functions √ Pythagorean theorem the height is s 3/2. The complementary angle has measure π/6, so we √ have cos π/6 = sin π/3 = 1/2, sin π/6 = cos π/3 = 3/2. The values of the other four basic trigonometric functions are determined as before. Identical to the first case, we can now find exact values of trigonometric functions for any angle which has reference angle π/6, or reference angle π/3. Example 1: Let’s find the exact value of some trigonometric functions at 4π/3. The reference angle is π/3 and P (4π/3) is in the third quadrant where both x and y coordinates are negative. √ Therefore cos 4π/3 = − cos π/3 = −1/2, and sin 4π/3 = − sin π/3 = − 3/2. Then for example √ tan 4π/3 = 3, since the ratio of two negative numbers is positive. Example 2: If we have a right triangle with base of length 12 and height 5, the Pythagorean theorem gives 13 as the length of the hypotenuse. In this case if θ denotes the angle adjacent to the base and hypotenuse, sin θ = 5/13, cos θ = 12/13, tan θ = 5/12, cot θ = 12/5, sec θ = 13/12 and csc θ = 13/5. Moreover we know how to find lots points on the unit circle which have a similar reference triangle. Example 3: If we have an isosceles right triangle of side 4, then if h denotes the length of the √ √ hypotenuse we have 4/h = sin(π/4) = 2/2. Here we write 8 = h 2 by cross multiplying and √ √ can solve for h = 8/ 2 = 4 2. It seems reasonable that we should be able to perform exercise 3 for any angle, θ, where 0 < θ < π/2. And of course it is possible, if we have inverse trigonometric functions. As with even power functions, due to their periodic nature trigonometric functions are not one-to-one. Therefore if we wish to invert them we must restrict their domain to a subset where they are one-to-one. In order to complete exercises like example 3 we want this restricted domain to include (0, π/2) and to be of most value it would be as large as possible. For y = sin x we use the convention that its domain is [−π/2, π/2] when discussing its inverse function denoted arcsin x. Formally the function is defined by the equations y = arcsin x iff x = sin y. Which means that we can draw a reference triangle with enclosed angle |y|, √ opposite side of length |x| and hypotenuse of length 1. The remaining side has length 1 − x2 by the Pythagorean theorem. This construct allows us to compute tan arcsin x = tan y = √ x/ 1 − x2 . When we start instead with y = cos x we use the convention that its domain is [0, π] and 34 Section 9: Special Values of Trigonometric Functions denote its inverse by arccos x. The conventional domain for y = tan x as an invertible function is (−π/2, π/2). Now we have y = arctan x iff x = tan y. So we can draw a reference triangle with enclosed angle |y|, opposite side of length |x| and adjacent side of length 1. The hypotenuse therefore has length √ √ x2 + 1, and we can write for example cos arctan x = cos y = 1/ x2 + 1. Notice that here the cosine function always has positive value which agrees with our convention about the domain of a composition of functions. One last picture to consider is the one flowing from considering sec x as an invertible function with inverse denoted arcsec x. Now we suppose that the domain of the secant function is limited to [0, π/2) ∪ (π/2, π]. By definition y = arcsec x iff x = sec y. So we can draw a right triangle with included angle |y|, hypotenuse of length x, adjacent side of length 1, and √ √ opposite side of length x2 − 1. Now for example csc arcsec x = csc y = x/ x2 − 1. The restricted domain for cot x as an invertible function is (0, π). For y = csc x to be considered invertible our convention is to limit its domain to [−π/2, 0) ∪ (0, π/2]. Exercises: 1. Find the exact value of each of the following: π 2π a) sin b) tan 3 3 −π 7π d) csc e) cos 6 6 3π g) sin (π) h) cot 4 j) cos (3π) k) tan (π) 2. Find the exact value of each of the following: a) tan arccos (−3/5) b) sin arccot (1/2) −5π c) cos 4 −π f) sin 3 3π i) sec 2 π l) csc 6 c) sec arccsc (2/3) 3. Sketch the picture described in the second paragraph of this section. Label the points O, P , B, D, and E. 4. Sketch and label a triangle in which: a) x = sin y b) x = tan y 35 Section 10: General Sinusoids A general sinusoid is a function of the form y = k + A sin(B(x − c)) or y = k + A cos(B(x − c)), where k, A, B and c are real numbers. Since the sine function and cosine function are horizontal shifts of each other we will only consider the form y = k + A sin(B(x − c)). Such a function is a shift right by c units (left if c < 0), of a compression horizontally by a factor of 1/B (expanded if B < 1), of a vertical scale by A units, of a shift up by k units (down if k < 0) of the basic function y = sin x. For such a function |A| is the amplitude. About the central line y = k the sine wave will fluctuate up |A| units and down |A| units. All function values are bounded below by k − |A| and above by k + |A|. The bounds are obtained when sin(B(x − c)) = ±1, which means that B(x − c) = (2n + 1)π/2, where n is an integer. This is equivalent to x = c + [(2n + 1)π/2B]. The general sinusoid is periodic with period 2π/B. π Example 1: Let’s sketch a graph of y = f (x) = 2 − 3 sin(3x − ). First we put the function 2 π in standard form y = f (x) = 2 − 3 sin(3(x − )). Then it’s a matter of determining where 6 f takes on maximal and minimal values, since half-way between it will take value 2. The extreme values happen when 3(x − π 6) = (2n + 1)π/2. So x − π 6 = (2n + 1)π/6. Thus 6x − π = (2n + 1)π = 2nπ + π. So 6x = (2n + 2)π, or x = (n + 1)π/3. We see that f (π/3) = −1 and f (2π/3) = 5. So f (π/2) = 2, and the rest is determined by symmetry. Of course we need to mark our x-axis off in units of π/6 to be able to clearly indicate where f takes on its extreme values and its central values. The y-axis needs to be marked so that all values between −1 and 5 can be seen. Elongating the x-axis a little gives a sketch which is smoother – more like the basic sine wave. We can also shorten the y-scale somewhat to acheive the same effect. 6 4 2 −π − 2π 3 π 3 − π3 -2 36 2π 3 π Section 10: General Sinusoids It is also useful to be able to combine two (or more) general sinusoids arithmetically. Given two sinusoids y = k1 + A1 sin(B1 (x − c1 )) and y = k2 + A2 sin(B2 (x − c2 )), their sum is is shift up k1 + k2 units of y = A1 sin(B1 (x − c1 )) + A2 sin(B2 (x − c2 )). This function is bounded since −(|A1 | + |A2 |) ≤ y ≤ |A1 | + |A2 |. The bounds may or may not be met. To graph this function we find scales for the x-axis and y-axis that allow both simpler functions to be graphed simultaneously. Then we add ordinates. Ordinate is a fancy name for the y-coordinate. Example 2: Sketch a graph of y = 3 sin x + sin(3x). First of all the outputs are bounded between −4 and 4. Second the function y = 3 sin x has period 2π, and is simply a vertical scale of the regular sine function. Third the function sin 3x has period 2π/3. So it is a horizontal compression of the sine wave by a factor of 3. To show 2 periods of each component function simultaneously we need an interval of x-values of length 4π. We might as well take x ∈ [−2π, 2π] since the overall function is odd. We do need to mark off multiples of π/6 on the x-axis in order to see all zeroes, and extreme values of sin 3x. We can also make two copies of our axes, one for scratch work, the other for the final product (neatness counts). 4 2 π −π −2π 2π -2 -4 Exercises: 1. For each of the following general sinusoidal functions identify the amplitude A, the vertical shift k, the period p, and the phase shift c. Then sketch and label a graph of the function on the given interval. In each instance choose scales for your axes which allow all important details to be clearly shown. At least two periods of the function should be shown. a) y = 2 − 3 cos(2πx + π ) 2 37 Section 10: General Sinusoids b) y = 1 + 2 sin( x π − ) 2 2 c) y = −3 + sin(3x) 2. Sketch and label a graph of each of the functions below on the specified interval. In each instance choose scales for your axes which allow all important details to be clearly shown. At least two periods of the resulting function should be graphed. a) y = 3 sin πx + 2 sin(2πx) b) y = 2 cos(2x) + cos(4x) 38 Section 11: Trigonometric Identities We’ve already seen many basic identities which trigonometric functions satisfy. For example if f is a basic trigonometric function, and g is its co-function, then f (x) = g(π/2−x), as long as both sides of the equation make sense. These are called the co-function identities. Another example are what are called the reciprocal identities, such as csc x = 1/ sin x. A third example is the quotient identity, tan x = sin x/ cos x. Finally, consider the Pythagorean identity sin2 x + cos2 x = 1. For each of these identities there is usually an equivalent way of stating the identity. For example we may rewrite the Pythagorean identity as sin2 x = 1−cos2 x, or as cos2 x = 1−sin2 x. We can even divide the equation through by cos2 x to obtain tan2 x+1 = sec2 x, after appealing to algebra, the quotient identity, and a reciprocal identity. This latest version of the identity is not true for all real x, but is true as long as cos2 x 6= 0, which means x 6= (2n+1)π/2. One game we might play is to continue generating equivalent identities from old using other identities. The result is a very impressive list of identities, too many for most people to memorize. So instead of recommending that people memorize every form of an identity, we recommend that they remember one form of the identity and develop the ability to re-derive the rest. This is very much like the difference between giving someone a fish, and teaching them how to fish. To build this skill a common practice is to use an exercise such as the following. Example 1: Verify that 1 + cot2 x = csc2 x is an identity. Solution : cos2 x sin2 x sin2 x + cos2 x = sin2 x 1 = sin2 x = csc2 x 1 + cot2 x = 1 + (quotient identity) (algebra) (Pythagorean identity) (reciprocal identity) Notice in our solution we did not start by affirming the conclusion. That is we did not suppose that the identity was true. We simply started with the left hand side and continued to apply identities until we arrived at the right hand side. Not only is this process logically proper, it is clearly reversible. So when we verify an identity we will use this process. Next 39 Section 11: Trigonometric Identities realize that every step was justified as either algebra or as the application of an identity. While this is not mandatory it is suggested. Third, observe that we did not show every single step of algebra. The real challenge is for each individual to learn to take as many steps as they need to verify the identity. Some people are more adroit than others combining algebraic steps. So two valid solutions might have a different number of delineated algebraic steps. Now to demonstrate the power of this process let’s derive one identity and develop a variety of other identities from it. Example 2: cos(θ − φ) = cos θ cos φ + sin θ sin φ, for θ, φ ∈ IR. Proof: Let θ, φ ∈ IR. On the unit circle the square of the distance from P (θ) to P (φ) is the same as the square of the distance from P (θ − φ) to P (0). Therefore by the distance formula we have (cos θ − cos φ)2 + (sin θ − sin φ)2 = (cos(θ − φ) − 1)2 + (sin(θ − φ) − 0)2 After expansion the left hand side is cos2 θ − 2 cos θ cos φ + cos2 φ + sin2 θ − 2 sin θ sin φ + sin2 φ. By applying the Pythagorean identity twice we get 2 − 2 cos θ cos φ − 2 sin θ sin φ. Meanwhile the right hand side becomes cos2 (θ −φ)−2 cos(θ −φ)+1+sin2 (θ −φ). Which after applying the Pythagorean identity once becomes 2 − 2 cos(θ − φ). Cancelling the 20 s and dividing through by −2 yields the identity. This is called the difference identity for cosine. From the difference identity for cosine we can derive the sum identity for cosine, Theorem: cos(θ + φ) = cos θ cos φ − sin θ sin φ. Proof: Here we simply write cos(θ + φ) = cos(θ − (−φ)) = cos θ cos(−φ) + sin θ sin(−φ) = cos θ cos φ − sin θ sin φ since cos(−φ) = cos φ and sin(−φ) = − sin φ. Similarly we can prove the sum identity for sine. Theorem: sin(θ + φ) = sin θ cos φ + sin φ cos θ 40 Section 11: Trigonometric Identities Proof: We use sin((π/2) − x) = cos x, sin(−z) = − sin z and cos(y − (π/2)) = sin y. π sin(θ + φ) = cos((θ + φ) − ) 2 π = cos(θ + (φ − )) 2 π π = cos θ cos(φ − ) − sin θ sin(φ − ) 2 2 π = cos θ sin φ + sin θ sin( − φ) 2 = cos θ sin φ + sin θ cos φ The rest is a matter of rearranging terms. The difference identity for sine is sin(θ − φ) = sin θ cos φ − sin φ cos θ. We leave the verification of this identity as an exercise. As interesting corollaries we get the double angle formulas which are 2 tan x cos 2x = cos2 x − sin2 x, sin 2x = 2 sin x cos x, and tan 2x = . For example 1 − tan2 x Theorem: cos 2x = cos2 x − sin2 x Proof: cos(2x) = cos(x + x) = cos x cos x − sin x sin x = cos2 x − sin2 x. Continuing, we can produce cos 2x = cos2 x − sin2 x = cos2 x + sin2 x − sin2 x − sin2 x = 1 − 2 sin2 x. This is an equivalent version of the double angle formula for cosine. One can also show that cos 2x = 2 cos2 x − 1. Each of the last two versions of the double angle formula for cosine can be rewritten. For 1 − cos 2x example sin2 x = . Which means that the function sin2 x can be analyzed via the 2 cosine function using shifting and scaling. Or if we wish we can derive one of the half-angle formulas. r x 1 − cos x Theorem: sin = ± 2 2 Proof: Set y = x/2, so 2y = x. Then sin2 (x/2) = sin2 y = (1 − cos 2y)/2 = (1 − cos x)/2. The result follows by taking square roots of both sides, realizing that the answer may be positive or negative. As a nice application of these formulas we have the following examples. Example 3: Find the exact value of sin(5π/12). 41 Section 11: Trigonometric Identities 8π 3π 2π π 5π = sin( − ) = sin( − ) 12 12 12 3 4 2π π π 2π = sin cos − sin cos 3 4 4 3 √ √ √ 3 2 21 − = 2 2 2 2 √ √ 6− 2 = 4 7π Example 4: Find the exact value of cos . 8 Solution: sin Solution: Since 7π/8 is in the second s quadrant s √ p √ 1 + cos 7π 1 + 22 7π 2+ 2 4 cos =− =− =− 8 2 2 2 Example 5: If we know that cos θ = 4/5 and 3π/2 ≤ θ ≤ 2π then cos 2θ = 2 cos2 θ − 1 = 2(16/25) − 1 = 7/25. While because θ is in quadrant four the Pythagorean identity implies sin θ = −3/5. So sin 2θ = 2 sin θ cos θ = −24/25. Therefore tan 2θ = −24/7, and we can also compute exact values for csc 2θ and cot 2θ. The previous example can be inverted by giving exact value of cos x, for example, and quadrantal information about x, to enable using the half-angle formulas to compute exact values of any of the trigonometric functions at x 2. Identities are also important since they may be used to simplify a problem. For example a3 if we want to analyze the formula z = 2 we put u = a tan θ. We find that a2 + u2 = (a + u2 )3/2 a2 + a2 tan2 θ = a2 (1 + tan2 θ) = a2 sec2 θ. So (a2 + u2 )3/2 = (a2 sec2 θ)3/2 = a3 sec3 θ, and z = cos3 θ. That is, trigonometric identities and composition of functions may allow us to change venue to a friendlier setting. To be sure no one has all trigonometric identities memorized, and many of them are not initially obvious. However we can for instance go one step further and write z = (cos 3θ + 3 cos θ)/4 which is a restatement of one of the homework exercises. This version of z as a function of θ can be visualized using addition of ordinates. Yet another application of trigonometric identities is in simplifying difference quotients of sin h trigonometric functions. For example by graphing we can observe that as h −→ 0, −→ 1, h cos h − 1 while −→ 0. We then have h sin(x + h) − sin x sin x cos h + sin h cos x − sin x = h h cos h − 1 sin h = sin x + cos x h h 42 Section 11: Trigonometric Identities So as h −→ 0, the difference quotient of the sine function approaches cos x. As a final demonstration of the power of these identities you may now read appendix E on rotation of axes to clear cross terms in general quadratic equations in two variables. Exercises: 1. Verify the following trigonometric identities. a) sin(θ − φ) = sin θ cos φ − sin φ cos θ tan θ + tan φ b) tan(θ + φ) = 1 − tan θ tan φ c) sin 2x = 2 sin x cos x 2 tan x d) tan 2x = 1 − tan2 x 1 + cos 2x e) cos2 x = 2 3 f) cos 3x = 4 cos x − 3 cos x r x 1 + cos x g) cos = ± 2 2 x 1 − cos x h) tan = 2 sin x 1 i) sin α cos β = (sin(α − β) + sin(α + β)) 2 1 j) cos α cos β = (cos(α − β) + cos(α + β)) 2 1 k) sin α sin β = (cos(α − β) − cos(α + β)) 2 2. Use trigonometric identities to find the exact value of each of the following. π 12 7π b) sec 12 5π c) csc 8 a) cot −4 and that θ is in quadrant III. Find the exact values of the 5 other five trigonometric functions at θ. 3 4. Given that θ is in quadrant I and that cos 2θ = − , find the exact values of the six 5 trigonometric functions at θ. 3. Using the fact that sin θ = 5. Read appendix E. 43 Section 12: Applications of Trigonometry Whenever a section in a math book has applications in the title it means “word exercises.” So now you know what you will mostly face in this section. The reason that many people have difficulty with word problems is that their solutions tend to be more involved than the “plug and chug” exercises otherwise found in text books. In fact the next to last part of a word problem usually is completing a particular type of “plug and chug” exercise. The difficulty is arriving at that stage. Here follows a short description of what is supposed to happen. Step 1: Read the problem carefully, several times. Make sure you understand what is being asked. Step 2: Draw a picture or diagram. Step 3: Label all known quantities including units, such as ft., lbs. etc. Usually word problems do not contain superfluous information. Thus any information given is vital. Step 4: Label all unknown quantities with distinct variables. In contrast to step 3, some unknown quantities may be irrelevant. With practice should come some facility in deciding which unknown quantities you can omit. Any unknown quantity which you are being asked to find must be labeled. Step 5: This is the hard part. Write down as many equations/relations/identities as you can which relate the known and unknown quantities. Some of these will be irrelevant. As before, with practice will come some ability to determine which equations are to be used. The hint is that the equations usually have something to do with the title of the section. Step 6: After eliminating as many variables as possible, hopefully you have reduced the word problem to a “plug and chug” exercise. Step 7: When you have solved the appropriate exercise, you should carefully and neatly re-write your solution eliminating all irrelevant unknown quantities,labels,equations,relations and identities. The finished product should look like the examples, possibly minus some of the prose. Before we get to the fun we need to develop a few more tools. First consider a triangle drawn with a horizontal base and one vertex labelled B above it. Label the left end point of the base A, the right end point C. Label the angle with vertex A, α, the angle with vertex B, β, and the angle with vertex C, γ. Finally label the length of the 44 Section 12: Applications of Trigonometry side opposite A, a, the length of the side opposite B, b, and the length of the third side c. Drop a perpendicular from B to the line the base lies in. Call the height of the perpendicular h, and the intersection of the perpendicular with the base P . If the perpendicular coincides with the side CB we have a right triangle and c2 = a2 + b2 . Otherwise the perpendicular splits the triangle into two right triangles AP B and BP C. The length from A to P is |c cos α|. Therefore the length from P to C is b − c cos α. Also we have c2 = h2 + c2 cos2 α and a2 = h2 + (b − c cos α)2 . From the first equation we have h2 = c2 − c2 cos2 α so by substitution a2 = h2 + (b − c cos α)2 = c2 − c2 cos2 α + b2 − 2bc cos α + c2 cos2 α = c2 + b2 − 2bc cos α This is the law of cosines. Since the parts of the triangle can be rearranged we also have b2 = a2 + c2 − 2ac cos β, and c2 = a2 + b2 − 2ab cos γ. Another tool arises from this scenario since in the triangles above we have h = a sin γ = sin α sin γ sin β c sin α. Therefore = . In fact whatever this common ratio is it must equal . a c b This is the law of sines. Though this law is simpler, and therefore more easy to use, it is more dangerous to use too. If we are only given a, c and α, then if a < c sin α, there is no triangle with these properties. If a = c sin α , then we get a unique right triangle where γ is the right angle. When a ≥ c(≥ c sin α) we again get only one triangle with these properties. For this triangle γ must be acute. However when c sin α < a < c, there are two angles γ1 and γ2 , where γ1 + γ2 = π and γ1 is obtuse while γ2 is acute. So when applying the law of sines we must be careful to consider whether our answer makes sense physically. Now for the fun. A fairly representative exercise whose solution uses trigonometry is the following. Example 1: An executive from a window washing company is putting in a bid for the weekly washing of the windows of a skyscraper. To set his price he needs to know approximately how tall the building is. His ex-wife works for the city planning office, so he doesn’t dare call them to ask for the information. Instead he moves away from the building 101 feet and measures the angle of elevation from that point to the roof of the building as 73◦ = 73π/180 radians. 45 Section 12: Applications of Trigonometry Thus the height of the building h satisfies h/101 = tan(73◦ ). So h is about 101 tan(73◦ ) ≈ 330 feet. Example 2: Fred and George are playing wallyball. George wants to pass the ball to Fred but a very tall opposing player is in the way. Fred is standing 4 feet away from a wall which George √ is standing 3 feet from. Fred and George are standing 145 feet apart and George decides that he wants to make a chest-high carom-pass off the wall to Fred. How far along the wall between them should George pass the ball. Solution: Let the point where George is standing be labelled A, and the point where Fred is standing B. Drop perpendiculars from A and B to the wall and label these points C and D respectively. Let P be the point on the wall which George should hit. Denote the length from P to C as x, so the length from P to D is 12 − x. Let θ be the angle with vertex P , initial side along the wall, and terminal side P A. Basic physics tells us that θ should also be the angle with x 12 − x vertex P , one side along the wall, and the other side P B. Therefore = cot θ = . Thus 3 4 1 4x = 3(12 − x) = 36 − 3x, which means x = 36/7 = 5 feet. By the way x/3 = 12/7 = cot θ, 7 so θ = arccot (12/7) ≈ 30.2◦ . Example 3: A car leaves a little town on a road traveling due South at 75 mph. Twenty minutes later a truck heads out of town traveling due East at 55 mph. How far apart in miles are the vehicles forty-five minutes after the truck departed? Solution: Let t = 0 correspond to when the truck departed and let us measure t in hours. Let a(t) denote the distance in miles of the car to the town as a function of t, and b(t) denote the dis75 tance of the truck in miles to the town as a function of time. Then a(t) = + 75t = 25 + 75t 3 and b(t) = 55t. By the Pythagorean theorem the distance between them in miles as a function of time t in hours is d(t) = p (25 + 75t)2 + (55t)2 = p p 625 + 3750t + 8650t2 = 5 25 + 150t + 346t2 p √ √ So d(3/4) = 5 25 + 150(3/4) + 346(9/16) = (5/4) 400 + 1800 + 3114 = (5/4) 5314 ≈ 91.1 miles. Example 4: A passenger jet leaves an airport at 12:00 pm traveling due west at 430 mph. A private jet leaves the same airport at 12:30 pm traveling 65◦ N of West at 552 mph. How far apart in miles are the planes at 2:00 pm? 46 Section 12: Applications of Trigonometry Solution: Let t denote time in hours and let t = 0 correspond to 12:30 pm. Let a(t) denote the distance of the passenger plane to the airport in miles. Let b(t) denote the distance of the private jet to the airport in miles. So similar to the previous example a(t) = 215 + 430t and b(t) = 552t. By the law of cosines the distance squared between them at time t is d2 (t) = (215 + 430t)2 + (552t)2 − 2(215 + 430t)(552t) cos(65◦ ). So when t = (3/2), the distance squared is d2 (3/2) = (4 · 215)2 + (828)2 − 2(860)(828) cos(65◦ ). Therefore p d(3/2) = (4 · 215)2 + (828)2 − 2(860)(828) cos(65◦ ) ≈ 907 miles. Example 5: A drug smuggler is traveling due North at 35 knots in a small boat. A coast guard cutter is 125 nautical miles away on a heading 22◦ East of North from the smuggler. What heading should the cutter take to intercept the smuggler, supposing that the cutter will run at its top speed of 40 knots? How long will it take? Solution: We could use the law of cosines here, but the law of sines is easier. Let A denote the position of the smuggler, P denote the position of the cutter and I denote the point of intercept. Let θ be the angle with vertex P and sides AP and IP . Let t denote time to intercept in hours. sin θ sin(22◦ ) Then by the law of sines = So sin θ = (35/40) sin(22◦ ) = (7/8) sin(22◦ ). Which 40t 35t means that θ = arcsin((7/8) sin(22◦ )) ≈ 19◦ . So the cutter should head roughly 49◦ South of sin(139◦ ) sin(22◦ ) West to intercept the smuggler. The time of intercept can be found using = 125 35t 125 sin(22◦ ) So t = ≈ 2.04 hours. 35 sin(139◦ ) Example 6: A pilot is trying to get to an airport 362 miles away which lies on a major highway that runs East and West. The control tower tells them to proceed on a heading 35◦ South of West. The pilot knows the compass in the plane is not reliable so they make a guess at the correct direction and proceed at 156 mph. Two hours later the pilot reaches the major highway. How far from the airport is she? Solution: Label the airport point A, the starting point of the plane P and the point where the airplane reaches the highway C. Then the length of AP is 362 miles, and the length of CP is 2 · 156 = 312 miles. The angle with vertex A and sides AP and AC has measure 35◦ . Since 362 sin(35◦ ) < 312 < 362 this is the ambiguous case of the law of sines. Still if we let γ denote the angle with vertex C and sides CA and CP , the law of sines implies sin γ = (362/312) sin(35◦ ) So γ = arcsin((181/156) sin(35◦ )) ≈ 41.7◦ or 138.3◦ . Unless the pilot has a completely lousy sense of direction, γ is the larger angle. Therefore the distance 47 Section 12: Applications of Trigonometry sin(35◦ ) sin(6.7◦ ) sin(6.7◦ ) = . Which means d = 312 ≈ 63.4 miles. If d 312 sin(35◦ ) we choose the other value of γ we find d ≈ 529.4 miles! d from A to P satisfies As an example of a possible non-word exercise we have Example 7: Show that for each A, B ∈ IR there is C, δ ∈ IR so that A sin x + B cos x = C sin(x + δ). Solution: Let C satisfy C 2 = A2 + B 2 , and δ = arctan(B/A). Draw a right reference triangle with sides of length |A|, |B| and |C| and included angle |δ|. Then A = C cos δ and B = C sin δ so that A sin x + B cos x = C cos δ sin x + C sin δ cos x = C sin(x + δ). Exercises: Give exact answers if possible. 1. The Butler Office Building (BOB) is 847 feet high. An apartment building stands across Main Avenue from BOB. Someone standing on the edge of the roof of the apartment building closest to BOB measures the angle of elevation to the top of BOB as 45◦ . They measure the angle of depression from their location to the foot of BOB as 60◦ . How far apart are the two buildings, and how tall is the apartment building? 2. An Army Reserve Squad is dropped off at point B and ordered to march on a heading of 42◦ North of East through a dense forest to a bivouac site at point A, 23 miles away on a North-South road. During the drop all their compasses are damaged. They take a guess at their proper direction and march off. Marching at 6 miles per hour they arrive at the road in 1 4 hours. How far are they from the bivouac site? How far off was their estimated heading? 2 3. Two jets depart LAX simultaneously at midnight. The first flies on a heading 32◦ North of East at 350 mph. The second flies on a heading 13◦ South of East at 467 mph. How far apart are the planes at 2:20 am? 4. Show that for each A, B ∈ IR there is C, δ ∈ IR so that A sin x + B cos x = C cos(x − δ) 5. John has a tree in his back yard that he wants removed. The tree removal company charges $25 a foot of height to remove it. John wants to know if he can afford it so he needs to know the height of the tree. He stands 21 feet away from the tree on level ground and measures the angle of elevation from himself to the top of the tree as 75◦ . About how much will it cost John to have the tree removed? 48 Section 13: Polar Coordinates A polar coordinate system is one in which there is a unique point in the plane called the pole and a ray through the pole called the polar axis. A point, P , in the plane can be coordinatized by its distance |r| to the pole and the measure of the angle θ from the polar axis to the ray through the pole and P . Normally we take the pole to lie at (0, 0) in the Cartesian coordinate system, and take the polar axis to be the positive x-axis. It is therefore common to measure θ in radians. Of course if θ and φ are coterminal angles, then the polar coordinates (r, θ) and (r, φ) describe the same point. So polar coordinates are not unique! In fact if we take the convention that when r is negative we move in the antipodal direction of θ, then the polar coordinates (r, θ) and (−r, θ ± π) represent the same point. To make our life easier we normally restrict the values of θ to lie in the range (−π, π]. With these facts in mind we seek a way to convert polar coordinates and Cartesian (or rectangular) coordinates. The point P (θ) lies on the line through the origin and (r, θ). Therefore by Euclid’s Theorem on similar triangles the Cartesian coordinates of the point are x = r cos θ, and y = r sin θ. Very simple! The conversion from rectangular coordinates to polar coordinates is not so smooth. When x = 0, the polar coordinates of the point are (y, π/2). When y = 0, the polar coordinates of the point are (x, 0). Otherwise x 6= 0 and the point with Cartesian coordinates (x, y) is a vertex of the nondegenerate right triangle whose other two vertices are (0, 0) and (x, 0). By p the Pythagorean theorem the distance from (x, y) to the origin is r = x2 + y 2 . Let θ denote the angle with vertex at the origin, initial side the positive x-axis and terminal side the line segment between the origin and (x, y). If α is the reference angle for this value of θ, then we know that tan α = |y|/|x| and in fact tan θ = y/x. So if −π/2 < θ < π/2, we can take θ = arctan(y/x). In general as long as θ is coterminal with an angle in the domain of the arctangent function, we have no problem. When θ is not coterminal with such and angle, θ + (2n + 1)π is for some integer n. We can therefore describe the point with rectangular coordinates (x, y), by the polar coordinates (−r, θ + (2n + 1)π). A major application of polar coordinates is that the description of certain graphs in polar coordinates lets us analyze them as functions, rather than relations. For example, x2 + y 2 = a2 has graph a circle with radius a centered at the origin. In polar coordinates x2 + y 2 = r 2 (squaring the top equation on the right) so we can describe this circle as r = a, a very simple 49 Section 13: Polar Coordinates formula in polar coordinates. By convention we do try to write polar equations in the form r = f (θ). So r is the dependent variable and θ is independent. In order to graph polar equations we use T-tables. π π π π The values we normally plug-in for θ are multiples of 0, , , , etc. That is values of θ 6 4 3 2 for which we can compute exact values for r. Standardly we keep the θ values in increasing order and after connecting our dots we use arrows to display the behavior of the curve as θ increases. This is called orienting the graph. This is a crucial part of the graph. If a > 0, the polar equations r = a and r = −a both represent the circle centered at the origin with radius a. The curve r = a is oriented positively, meaning that the motion along the curve as θ increases is counter-clockwise. The curve r = −a is oriented negatively, since the motion along the curve as θ increases is clockwise. These are therefore different curves in polar coordinates, though they transform to the same curve in rectangular coordinates. There is therefore some inherent danger when we attempt to convert equations in one coordinate system to the other. When we convert from polar to rectangular we lose orientation. When we convert from rectangular to polar we must often choose an orientation. The loose convention we use is to opt for the positively oriented polar curve. Example 1: Transform the rectangular equation y = 2x to a polar equation. Solution: By substitution we have r sin θ = 2r cos θ. So if r 6= 0, sin θ = 2 cos θ. Equivalently tan θ = 2, or θ = arctan 2. This works when r = 0 too, since we are at (0, θ) which is (0, 0) in rectangular coordinates. This kind of polar curve is not orientable, since θ is fixed. This is the polar equivalent of the vertical line in rectangular coordinates which has no slope. Example 2: Transform the polar equation r = 2 sin θ to a rectangular equation. Solution: Supposing that r 6= 0, we can multiply both sides by r to obtain r 2 = 2r sin θ. By substitution we have x2 + y 2 = 2y. By completing the square on y we get x2 + (y − 1)2 = 1. Thus this polar equation describes a circle of radius 1 tangent to the pole, oriented positively. Example 3: Transform the polar equation r = sin 2θ to a rectangular equation. Solution: Again supposing that r 6= 0 we first write r 3 = r 2 sin 2θ = r 2 (2 sin θ cos θ). Which we rewrite as r 3 = 2(r sin θ)(r cos θ). So by substitution (x2 + y 2 )3/2 = 2xy. Therefore (x2 + y 2 )3 = 4x2 y 2 . But unless you know about the classification of general expressions of total degree 6 in x and y, this equation will not help you sketch a graph of the curve. It does 50 Section 13: Polar Coordinates enable us to realize that the un-oriented curve is symmetric with respect to the y-axis, the x-axis, the line y = x, and with respect to the origin, since substituting in ±x for x and/or ±y for y leaves the equation unchanged. So if we figure out what the graph looks like in the first quadrant, we can apply symmetry to determine the rest. The values of θ we want to use to build a T-table are those for which we can compute exact values of sin 2θ. So for example 2θ = π/6, 2θ = π/4,, 2θ = π/3, etc. So we build the T-table θ 0 r(θ) 0 π/12 √ π/8 1/2 2/2 π/6 √ 3/2 π/4 1 Thus the graph of this polar equation is θ= π 3 θ= π 6 1 We don’t always need to transform a polar equation to rectangular form to investigate symmetry. If replacing θ with −θ leaves a polar equation unchanged, then its oriented graph will have polar axis (x-axis) symmetry. If replacing θ by π − θ leaves a polar equation unchanged, then the oriented graph will have symmetry with respect to the line θ = π/2 (the y-axis). Finally if replacing r by −r leaves a polar equation unchanged, then its oriented graph will have polar (original) symmetry. Example 3 indicates that these tests for polar symmetry are not as easy to implement as tests for rectangular symmetry. So there can be value to transforming polar equations to rectangular ones, even if the rectangular equations cannot be easily graphed. But the real problem persists, what does the graph of a polar equation look like? Aside from symmetry many polar equations have graphs which are periodic. This is because the most common functions to plug θ into are trigonometric. In such cases we are interested to 51 Section 13: Polar Coordinates know the range of θ values we must run through before the graph begins to repeat itself. In the example above, whereas sin 2θ has period π viewing it as a rectangular equation relating r and θ, the polar point (sin 2θ, θ) is antipodal to the polar point (sin(2(θ + π)), θ + π), since sin 2(θ + π) = sin(2θ + 2π) = sin 2θ. So in this case the period of the polar curve turns out to be 2π. Example 4: Sketch a graph of the polar equation r = sin 3θ. Solution: One can check that this graph will have y-axis symmetry. To determine further symmetries we plot r versus θ rectangularly. 1 −π − 2π 3 − π3 π 3 2π 3 π -1 From this we observe that the polar graph will be symmetric about θ = π/6, about θ = π/2 and about θ = 5π/6. Also we note that the period of the curve will be π since sin(3(θ + π)) = sin(3θ+3π) = sin 3θ cos 3π+sin 3π cos 3θ = − sin 3θ and therefore (sin 3θ, θ) and (− sin 3θ, θ+π) are the same point. Moreover the part of the polar curve about π/6 in shape will be the same as the part about π/2 with negative radii, and will be identical to that about 5π/6, back to positive radii. 52 Section 13: Polar Coordinates θ= π 3 θ= π 6 1 A general analysis leads one to recognize that a curve with polar equation r = a cos 2kθ, where k is a whole number, will have a graph which is a rose with 4k petals of length a, period 2π, and x-axis symmetry. The graph of r = a sin 2kθ, is virtually identical except that it has y-axis symmetry. In contrast the graph of the curve with polar equation r = a cos(2l + 1)θ, where l is a whole number, will be a rose with 2l + 1 petals of length a, each traversed twice as θ runs from 0 to 2π. This graph therefore has period π and also has x-axis symmetry. So rose curves might be analogous to quadratic equations in two variables, in that their behavior is completely describable by the form of the equation after it is put in standard form. Another class of polar curves with this property are the limaçons. A limaçon is a polar curve with equation of the form r = a + b cos θ, or r = a + b sin θ, where a and b are constants. The second class behave like the first except that they have y-axis symmetry instead of x-axis symmetry. For the first class there are three distinct possibilities: 1) a < b, 2) a = b, and 3) a > b. In each case the curve has period 2π. In the first case r = 0 twice in each period, since r = 0 implies that cos θ = −a/b < −1. Which means θ = arccos(−a/b), and 2π − arccos(−a/b). Also the sign of r changes at θ and again at 2π − θ. Consequently the graph has an inner loop. We distinguish this case by saying that the curve is a limaçon with an inner loop. In the second case r = 0 only once per period, when θ is an odd multiple of π. The result is what we called a cardioid. In the third case r is never zero and the graph looks rather like an egg, or a blunted ellipse. 53 Section 13: Polar Coordinates a<b a=b a>b Another application of polar coordinates is that they allow us to easily define and describe nondegenerate conic sections. This is a two-step process. First we define a conic as the set of points in the plane so that the distance from any point to a fixed point F , called the focus, is e times the distance from the point to a fixed line l, called the directrix. When e = 1 we have by definition a parabola. It turns out that e is exactly the eccentricity of the conic as discussed in section 7. We can then derive that if F is the pole, any conic has equation of the form r= ed , 1 + e cos θ r= ed , 1 − e cos θ r= ed , 1 + e sin θ or r= ed 1 − e sin θ where d is the distance from F to l. Exercises: 1. Plot the point with polar coordinates (5, 5π/6). b) Convert the polar coordinates to rectangular coordinates. c) Give two other pairs of polar coordinates which represent the same point, one pair with r > 0, the other with r < 0. 2. Convert the rectangular equation to a polar equation a) y = x b) x2 + y 2 = 4x c) y = x2 3. Sketch and label a graph of each polar equation clearly indicating orientation. a) r = 3 cos 3θ b) r = −3 cos 3θ c) r = 1 − 2 sin θ d) r = −1 − 2 sin θ e) r = 4 sin θ 4. Find a polar equation for the ellipse with focus at the pole, directrix with rectangular 1 equation x = 2 and eccentricity e = . 2 54 Section 14: Parametric Equations The last topic for us to treat is parametric equations, so called because both rectangular coordinates of a point on a curve are described by another quantity called the parameter. We normally take our third variable to be t and think of t as denoting time. This is the situation in many physical applications. A parametric curve is the set of all ordered pairs (x, y) where x = g(t) and y = h(t), where g(t) and h(t) describe x and y respectively in terms of t. An easy way to build a set of parametric equations is to start with y = f (x) and let x = g(t) = t be the parameter so that y = f (t) = h(t). Any function can be parametrized in this manner. There is a change in the graph of the curve in this case. As with polar curves, every parametric curve is oriented. We place arrow marks on the curve to denote the motion along the curve as the parameter t increases. Similarly any relation of the form x = k(y) can be parametrized by y. Every polar curve of the form r = f (θ) gives rise to a parametric curve parametrized by θ. The parametric equations are x = r cos θ = f (θ) cos θ = g(θ), and y = r sin θ = f (θ) sin θ = h(θ). This includes r = a, a circle with radius |a| centered at (0, 0). The standard parametric equations for this curve are x = a cos θ and y = a sin θ. This works because x2 + y 2 = a2 cos2 θ + a2 sin2 θ = a2 . This construction generalizes to x = h + a cos t, y = k + b sin t which are parametric equations for the ellipse (x − h)2 /a2 + (y − k)2 /b2 = 1. The sign of a and b don’t have to be positive so there are really four possible sets of parametric equations of this form for this ellipse x = h ± a cos t, and y = k ± b sin t. Since the crucial fact here is that sin2 t + cos2 t = 1, we can also parametrize this ellipse via x = h ± a sin t, y = k ± b cos t. These eight sets of parametric equations parametrize the same curve. They are distinct as is seen by considering which point corresponds to (x(0), y(0)) and whether the curve is oriented positively or negatively. If we tweak the Pythagorean identity to be in the form 1 = sec2 t − tan2 t, we can parametrize the hyperbola (x − h)2 /a2 − (y − k)2 /b2 = 1 via x = h ± a sec t, and y = k ± b tan t. In this case x = h ± a tan t and y = k ± b sec t parametrize the conjugate hyperbola (y − k)2 /b2 − (x − h)2 /a2 = 1. e2t + 2 + e−2t et − e−t et + e−t Define sinh t = and cosh t = , then cosh2 t = , and 2 2 4 e2t − 2 + e−2t sinh2 t = . So cosh2 t − sinh2 t = 1. We can therefore also use these functions, 4 which are the hyperbolic sine and cosine functions, to parametrize a hyperbola. 55 Section 14: Parametric Equations Since every parabola is either a function of the form y = k + (1/4c)(x − h)2 , or a relation of the form x = h + (1/4c)(y − k)2 every parabola is parametrizable. To summarize, every function is parametrizable, nearly every polar curve is parametrizable, and every conic section is parametrizable. As we see with ellipses and hyperbolas, there are often many ways to parametrize a curve. In fact if we have a set of parametric equations for a curve x = g(t) and y = h(t), then we can reparametrize part or all of the curve by s where t = f (s), and f is a function of s. The new parametric equations are x = g(t) = g(f (s)), and y = h(t) = h(f (s)). Example 1: Let x = t and y = x2 = t2 be a set of parametric equations describing the parabola y = x2 . If we set t = sin s, then x = t = sin s and y = t2 = sin2 s. Which means that the points on the curve x = sin s, y = sin2 s are on the curve y = x2 . However not every point of y = x2 is on this parametric curve since −1 ≤ x = sin s ≤ 1, and 0 ≤ sin2 s = y ≤ 1. This example is the first of a kind where we have the motion of the curve oscillating. If we want more or less of the curve we can use x = c sin u, and y = c2 sin u. This parametric curve describes the path of a particle that oscillates between (−c, c2 ) and (c, c2 ) along the path y = x2 . 12 10 8 6 4 2 -4 -2 2 4 x = 3 sin t, y = 9 sin2 t Example 2: Let C denote the part of the ellipse x2 /4 + y 2 /9 = 1 in the first quadrant. We can find sets of parametric equations for C in various ways. First we can solve for y in terms of x, since this part of the ellipse satisfies the horizontal line test. We get y 2 /9 = 1 − (x2 /4) = (4 − x2 )/4, so y 2 = (9/4)(4 − x2 ). Thus we can set x = t √ and y = (3/2) 4 − t2 , where t ∈ [0, 2]. 56 Section 14: Parametric Equations We can reverse orientation by setting x = t = 2 − s. Now 4 − t2 = 4 − (2 − s)2 = √ 4 − (4 − 4s + s2 ) = 4s − s2 , so y = (3/2) 4s − s2 , and s ∈ [0, 2]. √ We can increase speed as it were by setting x = t = 2s, and y = 3 1 − s2 , where s ∈ [0, 1]. √ We can decrease speed by setting x = t = (1/2)s and y = (3/4) 16 − s2 , where s ∈ [0, 4]. p √ We can call x = t = 2 sin θ, then y = (3/2) 4 − 4 sin2 θ = (3/2) 4 cos2 θ = 3 cos θ. Now we take θ ∈ [0, π/2]. Example 3: Let C be the line segment between (1, −2) and (3, 6). The line through these points has rectangular equation y = 4x − 6. We can therefore parametrize C by x = t, y = 4t − 6, where t ∈ [1, 3]. If we set t = s + 1, then x = s + 1, and y = 4(s + 1) − 6 = 4s − 2, where s ∈ [0, 2]. If we then set s = 2u we get x = 2u + 1, and y = 8u − 2, where u ∈ [0, 1]. If we set u = 1 − v we reverse the orientation and have x = 3 − 2v, and y = 6 − 8v, where v ∈ [0, 1]. If we have an object for which we know rectangular or polar equations, we can therefore parametrize the curve fairly easily. One challenge may be to find the best parameterization with respect to a given set of criteria. It might be instead that we want a parameterization with a specific orientation, and with a specified domain for the parameter. The tricks above work as long as we have at least one parameterization. A problem that arises is that we won’t always be given the rectangular or polar equations which a curve, or part of a curve satisfy. Instead we’ll begin with just a set of parametric equations. In order to plot the graph of a set of parametric equations we can use a three row/column table. The first row is labeled t and contains inputs. The second row consists of the outputs x = g(t). The third column consists of the outputs y = h(t). The ordered pairs (x, y) from the second and third columns belong on the graph. When we connect these dots it is again important to orient the graph to indicate the behavior as t increases. The number of values in the table varies from person to person. One should be careful to have a sufficient number of values. Also the t-values should be carefully chosen so that, if possible, exact values can be calculated for both x and y. Even when one is careful mistakes can easily be made. Eliminating the parameter to find an equation relating x and y is therefore a useful skill. We can then use what we know about this relation to ensure we’ve got a proper graph. 57 Section 14: Parametric Equations Example 4: Given the parametric equations x = tan θ + sec θ, and y = tan θ − sec θ, where −π/2 < θ < π/2 let us sketch a graph of the curve. The first thing we might do is to compute a three-row T-table where the input values of θ are ±π/6, ±π/4, ±π/3, and 0. θ −π/3 √ x(θ) 2 − √3 y(θ) −2 − 3 √−π/4 2−1 √ − 2−1 −π/6 √ 3/3 √ − 3 0 1 −1 π/6 √ √3 − 3/3 √π/4 2+1 √ − 2+1 π/3√ 2 + √3 −2 + 3 Plotting these points by hand and trying to connect the dots is dangerous. Much safer is to realize that xy = tan2 θ − sec2 θ = −1 by a variation of the Pythagorean identity. So the points of the parametric curve lie on the hyperbola xy = −1. 5 -5 5 -5 We may also need to remember that the parametric equations do not need to define the whole rectangular curve. √ Example 5: Given x = t and y = 1 − t, describe the curve parametrized. Solution: To begin with the domain for the parameter is at most [0, ∞). Secondly we can solve for t as a function of x to get t = x2 . So this set of parametric equations describes the right half of the downward opening parabola y = 1 − x2 . The orientation has (0, 1) = (x(0), y(0)) and as t increases the point moves toward (∞, −∞). Exercises: 1. Let x(t) = 1 + 4 sin t and y = −2 + 3 cos t. a) Find a rectangular equation for this relation. b) Sketch and label a graph of the curve clearly indicating orientation. 2. Let x(t) = t + 2 and y(t) = 3t − 5. a) Find a rectangular equation for this relation. b) Sketch and label a graph of the curve clearly indicating orientation. 58 Section 14: Parametric Equations 3. Find a set of parametric equations for the line through P = (1, −1) and Q = (−2, 4) so that P = (x(0), y(0)) and Q = (x(1), y(1)). 4. Find a set of parametric equations for the line through P = (2, 1) and Q = (−2, 2) so that Q = (x(1), y(1)) and P = (x(−1), y(−1)). 5. Find a set of parametric equations for the ellipse (x(0), y(0)) and (0, −4) = (x(π/2), y(π/2)). 59 x2 (y + 1)2 + = 1 so that (4, −1) = 16 9