3.6 applications in optimization

98
Applications in Optimization

Upload: math265

Post on 12-Apr-2017

651 views

Category:

Technology


2 download

TRANSCRIPT

Page 1: 3.6 applications in optimization

Applications in Optimization

Page 2: 3.6 applications in optimization

In this section, we solve various extrema problems in real world applications. These problems are referred to as “optimization problems” in mathematics, and derivatives play a major role in such problems.

Applications in Optimization

Page 3: 3.6 applications in optimization

extrema-existence–problem.

In this section, we solve various extrema problems in real world applications. These problems are referred to as “optimization problems” in mathematics, and derivatives play a major role in such problems. We begin with an

Applications in Optimization

Page 4: 3.6 applications in optimization

extrema-existence–problem. Let y be a function defined over an interval V, y may or may not have any extremum in V.

In this section, we solve various extrema problems in real world applications. These problems are referred to as “optimization problems” in mathematics, and derivatives play a major role in such problems. We begin with an

Applications in Optimization

Page 5: 3.6 applications in optimization

extrema-existence–problem. Let y be a function defined over an interval V, y may or may not have any extremum in V. Here are two examples of well behaved functions that don’t have extrema.

In this section, we solve various extrema problems in real world applications. These problems are referred to as “optimization problems” in mathematics, and derivatives play a major role in such problems. We begin with an

Applications in Optimization

Page 6: 3.6 applications in optimization

extrema-existence–problem. Let y be a function defined over an interval V, y may or may not have any extremum in V. Here are two examples of well behaved functions that don’t have extrema. Let y = tan(x) over the interval V = (–π/2, π/2) as shown.

In this section, we solve various extrema problems in real world applications. These problems are referred to as “optimization problems” in mathematics, and derivatives play a major role in such problems. We begin with an

Applications in Optimization

y=tan(x)

–π/2 π/2

x

y

Page 7: 3.6 applications in optimization

extrema-existence–problem. Let y be a function defined over an interval V, y may or may not have any extremum in V. Here are two examples of well behaved functions that don’t have extrema. Let y = tan(x) over the interval V = (–π/2, π/2) as shown. There is no extremum in V because V is open.

In this section, we solve various extrema problems in real world applications. These problems are referred to as “optimization problems” in mathematics, and derivatives play a major role in such problems. We begin with an

Applications in Optimization

y=tan(x)

–π/2 π/2

x

y

Page 8: 3.6 applications in optimization

Even if we close the interval V, it’s still not enough to guarantee the existence of an extremum.

Applications in Optimization

Page 9: 3.6 applications in optimization

Even if we close the interval V, it’s still not enough to guarantee the existence of an extremum.

Applications in Optimization

x

y=g(x)

For example, let y = g(x) be the function over the closed interval U = [–π/2, π/2] as shown.

–π/2 π/2

y (0, 1)

(0, –1)

Page 10: 3.6 applications in optimization

Even if we close the interval V, it’s still not enough to guarantee the existence of an extremum.

Applications in Optimization

x

y=g(x)

For example, let y = g(x) be the function over the closed interval U = [–π/2, π/2] as shown. Note that even though g(x) is bounded between

–π/2 π/2

y (0, 1)

(0, –1)

1 and –1, it still doesn’t have any extremum in U.

Page 11: 3.6 applications in optimization

Even if we close the interval V, it’s still not enough to guarantee the existence of an extremum.

Applications in Optimization

x

y=g(x)

For example, let y = g(x) be the function over the closed interval U = [–π/2, π/2] as shown. Note that even though g(x) is bounded between

–π/2 π/2

y (0, 1)

(0, –1)

1 and –1, it still doesn’t have any extremum in U.(Extrema Theorem for Continuous Functions)

Page 12: 3.6 applications in optimization

Even if we close the interval V, it’s still not enough to guarantee the existence of an extremum.

Applications in Optimization

x

y=g(x)

For example, let y = g(x) be the function over the closed interval U = [–π/2, π/2] as shown. Note that even though g(x) is bounded between

–π/2 π/2

y (0, 1)

(0, –1)

1 and –1, it still doesn’t have any extremum in U.(Extrema Theorem for Continuous Functions)Let y = f(x) be a continuous function defined over a closed interval V = [a, b], then both the absolute max. and the absolute min. exist in V.

Page 13: 3.6 applications in optimization

Even if we close the interval V, it’s still not enough to guarantee the existence of an extremum.

Applications in Optimization

x

y=g(x)

For example, let y = g(x) be the function over the closed interval U = [–π/2, π/2] as shown. Note that even though g(x) is bounded between

–π/2 π/2

y (0, 1)

(0, –1)

1 and –1, it still doesn’t have any extremum in U.(Extrema Theorem for Continuous Functions)Let y = f(x) be a continuous function defined over a closed interval V = [a, b], then both the absolute max. and the absolute min. exist in V. Furthermore, the absolute extrema must occur where f '(x) = 0,

Page 14: 3.6 applications in optimization

Even if we close the interval V, it’s still not enough to guarantee the existence of an extremum.

Applications in Optimization

x

y=g(x)

For example, let y = g(x) be the function over the closed interval U = [–π/2, π/2] as shown. Note that even though g(x) is bounded between

–π/2 π/2

y (0, 1)

(0, –1)

1 and –1, it still doesn’t have any extremum in U.(Extrema Theorem for Continuous Functions)Let y = f(x) be a continuous function defined over a closed interval V = [a, b], then both the absolute max. and the absolute min. exist in V. Furthermore, the absolute extrema must occur where f '(x) = 0, or where f'(x) is UDF, or they occur at the end points {a, b}.

Page 15: 3.6 applications in optimization

Applications in OptimizationHere are examples of each type of extrema.

Page 16: 3.6 applications in optimization

Applications in OptimizationHere are examples of each type of extrema.

A.

a xb c

Page 17: 3.6 applications in optimization

Applications in OptimizationHere are examples of each type of extrema.

A.

a xb

In A. The absolute maximum occurs at x = c where f '(c) = 0, and the absolute minimum occurs at the end point x = b.

c

A f'=0 ab. max.

An end–point ab. min.

Page 18: 3.6 applications in optimization

Applications in OptimizationHere are examples of each type of extrema.

A.

a xb

B.

a xb

In A. The absolute maximum occurs at x = c where f '(c) = 0, and the absolute minimum occurs at the end point x = b.

c

c d

A f'=0 ab. max.

An end–point ab. min.

Page 19: 3.6 applications in optimization

Applications in OptimizationHere are examples of each type of extrema.

A.

a xb

B.

a xb

In A. The absolute maximum occurs at x = c where f '(c) = 0, and the absolute minimum occurs at the end point x = b.

c

c d

In B. The absolute maximum occurs at x = c where f '(c) is UDF, and the absolute minimum occurs at x = dwhere f '(d) = 0.

A f'=0 ab. max.

An end–point ab. min.

A UDF– f' ab. max.

A f'=0 ab. min.

Page 20: 3.6 applications in optimization

Applications in OptimizationHere are examples of each type of extrema.

A.

a xb

B.

a xb

In A. The absolute maximum occurs at x = c where f '(c) = 0, and the absolute minimum occurs at the end point x = b.

c

c d

In B. The absolute maximum occurs at x = c where f '(c) is UDF, and the absolute minimum occurs at x = dwhere f '(d) = 0.

A f'=0 ab. max.

An end–point ab. min.

A UDF– f' ab. max.

A f'=0 ab. min.

Hence we have to consider points of all three cases when solving for extrema.

Page 21: 3.6 applications in optimization

Applications in OptimizationExample A. Let y be defined as the following.

x2 for 1 > xy =

2 – x for 1 ≤ x

Page 22: 3.6 applications in optimization

Applications in OptimizationExample A. Let y be defined as the following.

x2 for 1 > xy =

2 – x for 1 ≤ x(1, 1)

x

y

Page 23: 3.6 applications in optimization

Applications in OptimizationExample A. Let y be defined as the following.

x2 for 1 > xy =

2 – x for 1 ≤ x(1, 1)

x

y

Find the extrema of y over the interval [–1 , ½ ].

Page 24: 3.6 applications in optimization

Applications in OptimizationExample A. Let y be defined as the following.

x2 for 1 > xy =

2 – x for 1 ≤ x(1, 1)

x

y

Find the extrema of y over the interval [–1 , ½ ].The end–point values are f(–1) = 1 and f(½) = ¼.

Page 25: 3.6 applications in optimization

Applications in OptimizationExample A. Let y be defined as the following.

x2 for 1 > xy =

2 – x for 1 ≤ x(1, 1)

x

y

Find the extrema of y over the interval [–1 , ½ ].

x

y

The end–point values are f(–1) = 1 and f(½) = ¼.

–1 ½

(–1, 1)

(½, ¼ )

Page 26: 3.6 applications in optimization

Applications in OptimizationExample A. Let y be defined as the following.

x2 for 1 > xy =

2 – x for 1 ≤ x(1, 1)

x

y

Find the extrema of y over the interval [–1 , ½ ].

x

y

The end–point values are f(–1) = 1 and f(½) = ¼. y' = 0 at x = 0 so that (0, 0) is an f'=0–type minimum.

–1 ½

(–1, 1)

(½, ¼ )

Page 27: 3.6 applications in optimization

Applications in OptimizationExample A. Let y be defined as the following.

x2 for 1 > xy =

2 – x for 1 ≤ x(1, 1)

x

y

Find the extrema of y over the interval [–1 , ½ ].

x

y

Within the interval [–1 , ½ ], there is no point of the f'–UDF–type. Hence the ab. max. is at (–1, 1) and the ab. min is at (0, 0).

The end–point values are f(–1) = 1 and f(½) = ¼. y' = 0 at x = 0 so that (0, 0) is an f'=0–type minimum.

–1 ½

(–1, 1)

(½, ¼ )

Page 28: 3.6 applications in optimization

Applications in OptimizationThe key here is that if we are to find extrema of a given continuous function over a closed interval we have to obtain all three types of points described above, then select the correct solutions.

Page 29: 3.6 applications in optimization

Applications in Optimization

x

yFor instance, within [–½, 3], x = 0 is an f'=0–type solution.The point (1, 1) at x = 1 is an f'–UDF–type solution.

3–½

(1, 1)

The key here is that if we are to find extrema of a given continuous function over a closed interval we have to obtain all three types of points described above, then select the correct solutions.

Page 30: 3.6 applications in optimization

Applications in Optimization

x

yFor instance, within [–½, 3], x = 0 is an f'=0–type solution.The point (1, 1) at x = 1 is an f'–UDF–type solution. The end–point values are f(–½) = ¼ and f(3) = –1.

3–½

(1, 1)

The key here is that if we are to find extrema of a given continuous function over a closed interval we have to obtain all three types of points described above, then select the correct solutions.

(3, –1)

Page 31: 3.6 applications in optimization

Applications in Optimization

x

yFor instance, within [–½, 3], x = 0 is an f'=0–type solution.The point (1, 1) at x = 1 is an f'–UDF–type solution. The end–point values are f(–½) = ¼ and f(3) = –1.By inspection, the ab. max. is at (1, 1) and the ab. min is at (3, –1).

3–½

(1, 1)

The key here is that if we are to find extrema of a given continuous function over a closed interval we have to obtain all three types of points described above, then select the correct solutions.

(3, –1)

Page 32: 3.6 applications in optimization

Applications in Optimization

x

yFor instance, within [–½, 3], x = 0 is an f'=0–type solution.The point (1, 1) at x = 1 is an f'–UDF–type solution. The end–point values are f(–½) = ¼ and f(3) = –1.By inspection, the ab. max. is at (1, 1) and the ab. min is at (3, –1).

3–½

(1, 1)

The key here is that if we are to find extrema of a given continuous function over a closed interval we have to obtain all three types of points described above, then select the correct solutions.

(3, –1)

In the applied problems below, make sure all such relevant points are considered.

Page 33: 3.6 applications in optimization

Applications in OptimizationDistance Problems

Page 34: 3.6 applications in optimization

Applications in Optimization

The distance formula is a square–root formula whose radicand is always nonnegative. To find the extrema of a distance relation in the form of “√u(x)”, we drop the square–root and look for the extrema of u(x) instead.

Distance Problems

Page 35: 3.6 applications in optimization

Applications in Optimization

The distance formula is a square–root formula whose radicand is always nonnegative. To find the extrema of a distance relation in the form of “√u(x)”, we drop the square–root and look for the extrema of u(x) instead.

Distance Problems

Example B. Find the coordinate of the point on the line y = 2x that is closest to the point (1, 0).

Page 36: 3.6 applications in optimization

Applications in Optimization

The distance formula is a square–root formula whose radicand is always nonnegative. To find the extrema of a distance relation in the form of “√u(x)”, we drop the square–root and look for the extrema of u(x) instead.

Distance Problems

Example B. Find the coordinate of the point on the line y = 2x that is closest to the point (1, 0).

x

y=2x

(1, 0)

y

D

Page 37: 3.6 applications in optimization

Applications in Optimization

The distance formula is a square–root formula whose radicand is always nonnegative. To find the extrema of a distance relation in the form of “√u(x)”, we drop the square–root and look for the extrema of u(x) instead.

Distance Problems

Example B. Find the coordinate of the point on the line y = 2x that is closest to the point (1, 0).

x

y=2x

(1, 0)

y

(x, 2x)

A generic point on the line is (x, 2x) and its distance to the point (1, 0) is D = [(x – 1)2 + (2x – 0 )2]½.

D

Page 38: 3.6 applications in optimization

Applications in Optimization

The distance formula is a square–root formula whose radicand is always nonnegative. To find the extrema of a distance relation in the form of “√u(x)”, we drop the square–root and look for the extrema of u(x) instead.

Distance Problems

Example B. Find the coordinate of the point on the line y = 2x that is closest to the point (1, 0).

x

y=2x

(1, 0)

y

(x, 2x)

A generic point on the line is (x, 2x) and its distance to the point (1, 0) is D = [(x – 1)2 + (2x – 0 )2]½. We want to find the x value that gives the minimal D. Note that D is defined for all numbers.

D

Page 39: 3.6 applications in optimization

Applications in Optimization

x

(1, 0)

(x, 2x)

We minimize the radicandu = D2 = (x – 1)2 + (2x – 0 )2 = 5x2 – 2x + 1 instead.

y

D

Page 40: 3.6 applications in optimization

Applications in Optimization

x

(1, 0)

(x, 2x)

We minimize the radicandu = D2 = (x – 1)2 + (2x – 0 )2 = 5x2 – 2x + 1 instead. Set u' = 10x – 2 = 0. We get x = 1/5, and the point is (1/5, 2/5).

y

D

Page 41: 3.6 applications in optimization

Applications in Optimization

x

(1, 0)

(x, 2x)

We minimize the radicandu = D2 = (x – 1)2 + (2x – 0 )2 = 5x2 – 2x + 1 instead. Set u' = 10x – 2 = 0. We get x = 1/5, and the point is (1/5, 2/5).From the geometry it’s obvious (1/5, 2/5) is the closest point to (1, 0).

y

D

Page 42: 3.6 applications in optimization

Applications in Optimization

x

(1, 0)

(x, 2x)

We minimize the radicandu = D2 = (x – 1)2 + (2x – 0 )2 = 5x2 – 2x + 1 instead. Set u' = 10x – 2 = 0. We get x = 1/5, and the point is (1/5, 2/5).From the geometry it’s obvious (1/5, 2/5) is the closest point to (1, 0).

½

y

D

Similarly if we wish to find the extrema of eu(x) or In(u(x)), we may find the extrema of u(x) instead. This is true because like sqrt(x), eu(x) and In(u(x)) are increasing functions, i.e. if s < t, then es < et,so the extrema of eu(x) are the same as u(x).

Page 43: 3.6 applications in optimization

Applications in Optimization

x

(1, 0)

(x, 2x)

We minimize the radicandu = D2 = (x – 1)2 + (2x – 0 )2 = 5x2 – 2x + 1 instead. Set u' = 10x – 2 = 0. We get x = 1/5, and the point is (1/5, 2/5).From the geometry it’s obvious (1/5, 2/5) is the closest point to (1, 0).

½

y

D

Similarly if we wish to find the extrema of eu(x) or In(u(x)), we may find the extrema of u(x) instead. This is true because like sqrt(x), eu(x) and In(u(x)) are increasing functions, i.e. if s < t, then es < et,so the extrema of eu(x) are the same as u(x). In fact, if f(x) is an increasing function, then the extrema of f(u(x)) are the same as the ones of u(x).

Page 44: 3.6 applications in optimization

Applications in Optimization(What is the corresponding statement about the extrema of f(u(x)) and u(x) if f(x) is a decreasing function? Hint, consider f(x) = –x and f(u(x)) = –u(x).)

Page 45: 3.6 applications in optimization

Applications in Optimization(What is the corresponding statement about the extrema of f(u(x)) and u(x) if f(x) is a decreasing function? Hint, consider f(x) = –x and f(u(x)) = –u(x).)Area ProblemsIn area optimization problems, the main task is to express the area in question in a single well chosen variable x as A(x).

Page 46: 3.6 applications in optimization

Applications in Optimization(What is the corresponding statement about the extrema of f(u(x)) and u(x) if f(x) is a decreasing function? Hint, consider f(x) = –x and f(u(x)) = –u(x).)Area ProblemsIn area optimization problems, the main task is to express the area in question in a single well chosen variable x as A(x). If x represents a specific measurement, as opposed to a coordinate, then x must be nonnegative (often it’s also bounded above).

Page 47: 3.6 applications in optimization

Applications in Optimization(What is the corresponding statement about the extrema of f(u(x)) and u(x) if f(x) is a decreasing function? Hint, consider f(x) = –x and f(u(x)) = –u(x).)Area ProblemsIn area optimization problems, the main task is to express the area in question in a single well chosen variable x as A(x). If x represents a specific measurement, as opposed to a coordinate, then x must be nonnegative (often it’s also bounded above). In such cases consider the significance when x = 0 (or it’s other boundary value). Often one of the extrema is at the boundary, and we are looking for the other extremum. A drawing is indispensable in any geometric problem.

Page 48: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible? x

x/2 y

Page 49: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible?The enclosed area is A = xy.

x

x/2 y

Page 50: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible?The enclosed area is A = xy. The total linear lengthis 120, that is, 3y + 2½ x = 120

x

x/2 y

Page 51: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible?The enclosed area is A = xy. The total linear lengthis 120, that is, 3y + 2½ x = 120 or that y = 40 – 5x/6.

x

x/2 y

Page 52: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible?The enclosed area is A = xy. The total linear lengthis 120, that is, 3y + 2½ x = 120 or that y = 40 – 5x/6. So A(x) = x(40 – 5x/6) = 40x– 5x2/6.

x

x/2 y

Page 53: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible?The enclosed area is A = xy. The total linear lengthis 120, that is, 3y + 2½ x = 120 or that y = 40 – 5x/6. So A(x) = x(40 – 5x/6) = 40x– 5x2/6. At the boundary values x = 0, 48 (why?) we have the absolute minimal A = 0.

x

x/2 y

Page 54: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible?The enclosed area is A = xy. The total linear lengthis 120, that is, 3y + 2½ x = 120 or that y = 40 – 5x/6. So A(x) = x(40 – 5x/6) = 40x– 5x2/6. At the boundary values x = 0, 48 (why?) we have the absolute minimal A = 0. Setting A'(x) = 40 – 5x/3 = 0, we get x = 24

x

x/2 y

Page 55: 3.6 applications in optimization

Applications in OptimizationExample C. Farmer Fred wants to raise chickens, ducks and pigs. He has 120 yards of fence to build rectangular regions as shown to separate the animals. What is the maximal enclosed area that is possible?The enclosed area is A = xy. The total linear lengthis 120, that is, 3y + 2½ x = 120 or that y = 40 – 5x/6. So A(x) = x(40 – 5x/6) = 40x– 5x2/6. At the boundary values x = 0, 48 (why?) we have the absolute minimal A = 0. Setting A'(x) = 40 – 5x/3 = 0, we get x = 24 which yields y = 20. So 20×24 = 480 yd2 must be the maximum area.

x

x/2 y

Page 56: 3.6 applications in optimization

Applications in OptimizationIn fact for any specified partition of the rectangle, the maximal area always occurs when the total–vertical–fence–length is equal to the total–horizontal–fence–length. That is, using half of the fence for vertical dividers and the other half for the horizontal dividers will give the maximum enclosed area.Hence if farmer Joe is to build an enclosure with 120 yd of fence as shown, then the answer is x = 60/3½ = 120/7, y = 60/3 = 20 with the maximum possible area of 120/7×20 ≈ 343 yd2. x

x/2

y

Page 57: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

Page 58: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

The volume of a cylinder is V = Bh where B = area of the base and h = height

Page 59: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

The volume of a cylinder is V = Bh where B = area of the base and h = height so we have120π = πr2hor h = 120/r2.

Page 60: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

The volume of a cylinder is V = Bh where B = area of the base and h = height so we have120π = πr2hor h = 120/r2. The surface area of the cylinder consists of a circularbase and the circular wall.

Page 61: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

The volume of a cylinder is V = Bh where B = area of the base and h = height so we have120π = πr2hor h = 120/r2. The surface area of the cylinder consists of a circularbase and the circular wall. The base area is πr2.

Page 62: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

The volume of a cylinder is V = Bh where B = area of the base and h = height so we have120π = πr2hor h = 120/r2. The surface area of the cylinder consists of a circularbase and the circular wall. The base area is πr2. Unroll and flatten the wall, it is an h x 2πr rectangular sheet.

Page 63: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

The volume of a cylinder is V = Bh where B = area of the base and h = height so we have120π = πr2hor h = 120/r2. The surface area of the cylinder consists of a circularbase and the circular wall. The base area is πr2. Unroll and flatten the wall, it is an h x 2πr rectangular sheet. Hence the total surface is S = πr2+ 2πrh,or S = πr2 + 2πr( )120

r2

Page 64: 3.6 applications in optimization

Applications in OptimizationExample D. Farmer Joe wishes to build an open cylindrical water tank made from sheet–metal that will hold 120π yd3 of water. What is the least amount of sheet–metal (in yd2) needed?

120 yd3

h

r

The volume of a cylinder is V = Bh where B = area of the base and h = height so we have120π = πr2hor h = 120/r2. The surface area of the cylinder consists of a circularbase and the circular wall. The base area is πr2. Unroll and flatten the wall, it is an h x 2πr rectangular sheet. Hence the total surface is S = πr2+ 2πrh,or S = πr2 + 2πr( ) = πr2 + 240π/r = π(r2 + 240/r)120

r2

Page 65: 3.6 applications in optimization

Applications in Optimization

120 yd3

h

r

Setting S' = 2π(r – 120/r2) = 0,dividing both sides by 2π and clearing the denominator yields r3 – 120 = 0

Page 66: 3.6 applications in optimization

Applications in Optimization

120 yd3

h

r

Setting S' = 2π(r – 120/r2) = 0,dividing both sides by 2π and clearing the denominator yields r3 – 120 = 0

or that r = 1201/3, which gives the absolute minimal (why?) surface

Page 67: 3.6 applications in optimization

Applications in Optimization

120 yd3

h

r

Setting S' = 2π(r – 120/r2) = 0,dividing both sides by 2π and clearing the denominator yields r3 – 120 = 0

or that r = 1201/3, which gives the absolute minimal (why?) surface

of π(1202/3 + ) = π 360/1201/3 2401201/3 yd2.

Page 68: 3.6 applications in optimization

Applications in Optimization

120 yd3

h

r

Setting S' = 2π(r – 120/r2) = 0,dividing both sides by 2π and clearing the denominator yields r3 – 120 = 0

or that r = 1201/3, which gives the absolute minimal (why?) surface

of π(1202/3 + ) = π 360/1201/3 2401201/3 yd2.

If the geometric object has a variable angle θ, then θ may be utilized as the independent variable.

Page 69: 3.6 applications in optimization

Applications in Optimization

120 yd3

h

r

Setting S' = 2π(r – 120/r2) = 0,dividing both sides by 2π and clearing the denominator yields r3 – 120 = 0

or that r = 1201/3, which gives the absolute minimal (why?) surface

of π(1202/3 + ) = π 360/1201/3 2401201/3 yd2.

If the geometric object has a variable angle θ, then θ may be utilized as the independent variable.Example E. Find the isosceles triangle having two sides of length 4 with the largest area as shown.

44

Page 70: 3.6 applications in optimization

Applications in OptimizationThere are two observations that will make the algebra cleaner. 4

4

Page 71: 3.6 applications in optimization

Applications in OptimizationThere are two observations that will make the algebra cleaner. First let’s take advantage of the symmetry. Draw a perpendicular line as shown, it cuts the triangle

44

into two mirror image right triangles.

Page 72: 3.6 applications in optimization

Applications in OptimizationThere are two observations that will make the algebra cleaner. First let’s take advantage of the symmetry. Draw a perpendicular line as shown, it cuts the triangle

44

into two mirror image right triangles. We may maximize one of the right triangles instead, i.e. to find the largest area possible of a right triangle with hypotenuse equal to 4.

4

a

bA

Page 73: 3.6 applications in optimization

Applications in OptimizationThere are two observations that will make the algebra cleaner. First let’s take advantage of the symmetry. Draw a perpendicular line as shown, it cuts the triangle

44

into two mirror image right triangles. We may maximize one of the right triangles instead, i.e. to find the largest area possible of a right triangle with hypotenuse equal to 4.

4

a

b

Second, instead of expressing the area A in terms of the measurements of the legs a and b, we express A in terms of the angle θ as shown.

A

θ

Page 74: 3.6 applications in optimization

Applications in OptimizationA = ½ ab so in terms of θ we have

4

a

bA

θ

A = ½ 4sin(θ) 4cos(θ) = 8 sin(θ) cos(θ)

Page 75: 3.6 applications in optimization

Applications in OptimizationA = ½ ab so in terms of θ we have

Note that 0 ≤ θ ≤ π/2.

4

a

bA

θ

A = ½ 4sin(θ) 4cos(θ) = 8 sin(θ) cos(θ)

Page 76: 3.6 applications in optimization

Applications in OptimizationA = ½ ab so in terms of θ we have

Note that 0 ≤ θ ≤ π/2.Set A'(θ) = 8(cos2(θ) – sin2 (θ)) = 0

4

a

bA

θ

A = ½ 4sin(θ) 4cos(θ) = 8 sin(θ) cos(θ)

Page 77: 3.6 applications in optimization

Applications in OptimizationA = ½ ab so in terms of θ we have

Note that 0 ≤ θ ≤ π/2.Set A'(θ) = 8(cos2(θ) – sin2 (θ)) = 0

4

a

bA

θ

A = ½ 4sin(θ) 4cos(θ) = 8 sin(θ) cos(θ)

However, the only solution with 0 ≤ θ ≤ π/2 is θ = π/4Hence cos(θ) = ±sin (θ)

Page 78: 3.6 applications in optimization

Applications in OptimizationA = ½ ab so in terms of θ we have

Note that 0 ≤ θ ≤ π/2.Set A'(θ) = 8(cos2(θ) – sin2 (θ)) = 0

4

a

b

which must be a maximum (why?). So the largest possible isosceles triangle in question is the right isosceles triangle as shown with area 2A = ½ (4 ×4) = 8.

A

θ

A = ½ 4sin(θ) 4cos(θ) = 8 sin(θ) cos(θ)

However, the only solution with 0 ≤ θ ≤ π/2 is θ = π/4Hence cos(θ) = ±sin (θ)

44 π/2

Page 79: 3.6 applications in optimization

Applications in OptimizationExample F. A duck can walk 3 ft/sec and swim2 ft/sec at its top speed. It sees a piece of bread across the pool as shown. What is the path it should take to get to the bread as soon as possible?

40 ft

30 ft

pool

Page 80: 3.6 applications in optimization

Applications in OptimizationExample F. A duck can walk 3 ft/sec and swim2 ft/sec at its top speed. It sees a piece of bread across the pool as shown. What is the path it should take to get to the bread as soon as possible?

40 ft

30 ft

poolSelect the x as shown in the picture.

x

Page 81: 3.6 applications in optimization

Applications in OptimizationExample F. A duck can walk 3 ft/sec and swim2 ft/sec at its top speed. It sees a piece of bread across the pool as shown. What is the path it should take to get to the bread as soon as possible?

40 ft

30 ft

poolSelect the x as shown in the picture. (The selection of the correct x is important. The wrong choice leads to tangled algebra as is often the case in these problems.)

x

Page 82: 3.6 applications in optimization

Applications in OptimizationExample F. A duck can walk 3 ft/sec and swim2 ft/sec at its top speed. It sees a piece of bread across the pool as shown. What is the path it should take to get to the bread as soon as possible?

40 ft

30 ft

poolSelect the x as shown in the picture. (The selection of the correct x is important. The wrong choice leads to tangled algebra as is often the case in these problems.)

Hence if your choice of x leads to nowhere, try another choice.

x

Page 83: 3.6 applications in optimization

Applications in OptimizationExample F. A duck can walk 3 ft/sec and swim2 ft/sec at its top speed. It sees a piece of bread across the pool as shown. What is the path it should take to get to the bread as soon as possible?

40 ft

30 ft

pool

x

Note that 0 ≤ x ≤ 40, where x = 0 means the duck walks aroundthe pool and x = 40 means the duck swims all the way.

Select the x as shown in the picture. (The selection of the correct x is important. The wrong choice leads to tangled algebra as is often the case in these problems.)

Page 84: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x(x2 + 900)½

For x ≠ 0, the duck swims a distance of (x2 + 900)½

Page 85: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x(x2 + 900)½

40 – x

For x ≠ 0, the duck swims a distance of (x2 + 900)½ and walks a distance of (40 – x),

Page 86: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x

and the total time t it takes is(x2 + 900)½

40 – xt(x) = (x2 + 900)½

2 +(40 – x)

3

For x ≠ 0, the duck swims a distance of (x2 + 900)½ and walks a distance of (40 – x),

swimming time walking time

Page 87: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x

and the total time t it takes is(x2 + 900)½

40 – xt(x) = (x2 + 900)½

2 +(40 – x)

3Set t'(x) = 2(x2 + 900)½

x – 31

For x ≠ 0, the duck swims a distance of (x2 + 900)½ and walks a distance of (40 – x),

= 0

Page 88: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x

and the total time t it takes is(x2 + 900)½

40 – xt(x) = (x2 + 900)½

2 +(40 – x)

3Set t'(x) = 2(x2 + 900)½

x – 31

For x ≠ 0, the duck swims a distance of (x2 + 900)½ and walks a distance of (40 – x),

= 0

We have x = 31

2(x2 + 900)½

Page 89: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x

and the total time t it takes is(x2 + 900)½

40 – xt(x) = (x2 + 900)½

2 +(40 – x)

3Set t'(x) = 2(x2 + 900)½

x – 31

For x ≠ 0, the duck swims a distance of (x2 + 900)½ and walks a distance of (40 – x),

= 0

We have x = 31

2(x2 + 900)½ 3x = 2(x2 + 900)½

Page 90: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x

and the total time t it takes is(x2 + 900)½

40 – xt(x) = (x2 + 900)½

2 +(40 – x)

3Set t'(x) = 2(x2 + 900)½

x – 31

For x ≠ 0, the duck swims a distance of (x2 + 900)½ and walks a distance of (40 – x),

= 0

We have x = 31

2(x2 + 900)½ 3x = 2(x2 + 900)½ 9x2 = 4(x2 + 900)

Page 91: 3.6 applications in optimization

Applications in Optimization

40 ft

30 ft

pool

x

and the total time t it takes is(x2 + 900)½

40 – xt(x) = (x2 + 900)½

2 +(40 – x)

3Set t'(x) = 2(x2 + 900)½

x – 31

For x ≠ 0, the duck swims a distance of (x2 + 900)½ and walks a distance of (40 – x),

= 0

We have x = 31

2(x2 + 900)½ 3x = 2(x2 + 900)½ 9x2 = 4(x2 + 900)

5x2 = 3600 x = ±12√5Discarding the negative answer, x =12√5 ≈ 26.8 ft.

Page 92: 3.6 applications in optimization

Applications in OptimizationWhen x = 0, the time it takes is not t(0) because t(x) assume the duck swims at least 30 ft,

Page 93: 3.6 applications in optimization

Applications in OptimizationWhen x = 0, the time it takes is not t(0) because t(x) assume the duck swims at least 30 ft, as it would be the case if the duck is across a river from the bread instead of a pool.

Page 94: 3.6 applications in optimization

Applications in OptimizationWhen x = 0, the time it takes is not t(0) because t(x) assume the duck swims at least 30 ft, as it would be the case if the duck is across a river from the bread instead of a pool. Let’s examine the three points.

Page 95: 3.6 applications in optimization

Applications in Optimization

When x = 40 the duck swims all the way and it takes t(40) = 50/2 = 25 seconds.

When x = 0, the time it takes is not t(0) because t(x) assume the duck swims at least 30 ft, as it would be the case if the duck is across a river from the bread instead of a pool. Let’s examine the three points. When x = 0, the duck walks 70 ft and it takes 70/3 = 23 1/3 seconds.

At the critical point, t(12√5) ≈ 24.5 seconds.

Page 96: 3.6 applications in optimization

Applications in Optimization

When x = 40 the duck swims all the way and it takes t(40) = 50/2 = 25 seconds.

When x = 0, the time it takes is not t(0) because t(x) assume the duck swims at least 30 ft, as it would be the case if the duck is across a river from the bread instead of a pool. Let’s examine the three points. When x = 0, the duck walks 70 ft and it takes 70/3 = 23 1/3 seconds.

At the critical point, t(12√5) ≈ 24.5 seconds.Therefore the duck should walk around the pool.

Page 97: 3.6 applications in optimization

Applications in Optimization

When x = 40 the duck swims all the way and it takes t(40) = 50/2 = 25 seconds.

When x = 0, the time it takes is not t(0) because t(x) assume the duck swims at least 30 ft, as it would be the case if the duck is across a river from the bread instead of a pool. Let’s examine the three points. When x = 0, the duck walks 70 ft and it takes 70/3 = 23 1/3 seconds.

At the critical point, t(12√5) ≈ 24.5 seconds.Therefore the duck should walk around the pool.Remarks1. As noted that if the duck is across a river instead of a pool, then indeed x = 12√5 is the absolute minimum because t(0) = swim + walk = 28 1/3 sec.

Page 98: 3.6 applications in optimization

Applications in Optimization2. Just because the duck walks faster does not automatically means it should always walk around the pool. We may think about this by imagining the duck is slightly faster in walking, say at 2.01 ft/sec,than swimming at 2 ft/sec. Then of course the duck should swim toward the bread instead of wasting time walking around. In fact the closer the walking speed is to the swimming speed, the more the duck should swim toward the bread.

Back to math–265 pg