Business Mathematics Homework #9 Due: ID: Name:

A. Multi-variable optimization
1. Find the stationary points and classify them when
f ( x, y ) = x 3 − x 2 − y 2 + 8.
Apply Theorem 13.3.1.

Theorem 13.3.1
Suppose f ( x, y ) is a C 2 function in a domain S , and let ( x0 , y0 ) be
an interior stationary point of S . Write
= A f= ′′
11 ( x0 , y0 ), B f12′′ ( x0 ,=
y0 ), and C f 22′′ ( x0 , y0 )
(a) If A < 0 and AC − B 2 > 0, then ( x0 , y0 ) is a (strict) local maximum point.
(b) If A > 0 and AC − B 2 > 0, then ( x0 , y0 ) is a (strict) local minimum point.
(c) If AC − B 2 < 0, then ( x0 , y0 ) is a saddle point.
(d) If AC − B 2 =
0, then ( x0 , y0 ) could be a local maximum,
a local minimum, or a saddle point.
Solution.
 f ( x, y ) = x 3 − x 2 − y 2 + 8.
 The stationary points must satisfy the two equations
f1′( x, y ) = 0 and f 2′( x, y ) =
3x 2 − 2 x = −2 y =
0
 Because 3 x 2 − 2 x= x ( 3 x − 2 ) , we see that the first equation has
the solutions x = 0 and x = 2 3.
 The second equation has the solution y = 0.
 We conclude that (0, 0) and (2 3, 0) are the only stationary points.
 Furthermore, f11′′ ( x, y ) =
6 x − 2, f12′′ ( x, y ) =
0, and f 22′′ ( x, y ) =
−2.
 A convenient way of classifying the stationary points is to make a table
like the following (with A, B, and C defined in Theorem 13.3.1):
( x, y ) A B C AC − B 2 Type of point
(0, 0) −2 0 −2 4 Local maximum point
(2 3, 0) 2 0 −2 −4 Saddle point
2. A firm sells a product in two markets over which it has monopolistic power.
The firm has two independent demand curves:
P1 =
1 − 3Q1 , P2 =
2 − 4Q2
where P1 and P2 are prices in markets 1 and 2, respectively,
and Q1 and Q2 are production (sales) amounts in markets 1 and 2, respectively.
The total cost is proportional to total production:
(Q) 5 ( Q1 + Q2 )
C=
What are the optimal prices of ( P1 ,P2 ) in markets 1 and 2?
Solution.
 In general, we let
P1 =
a1 − b1Q1 , P2 =
a2 − b2Q2
where=
a1 1,=
b1 3,=
a2 2 and=
b2 4.
α ( Q1 + Q2 ) where α =
 C (Q) = 5
 As a function of Q1 and Q2 , total profits are
1 1 + P2 Q2 − α ( Q1 + Q2 )
π (Q1 , Q2 ) = PQ
= ( a1 − b1Q1 ) Q1 + ( a2 − b2Q2 ) Q2 − α ( Q1 + Q2 )
= ( a1 − α ) Q1 + ( a2 − α ) Q2 − b1Q12 − b2Q22
 We want to find the values of Q1 ≥ 0 and Q2 ≥ 0 that maximize profits.
 The first-order conditions are
π 1′(Q1 , Q2 ) = ( a1 − α ) − 2b1Q1 = 0, π 2′ (Q1 , Q2 ) = ( a2 − α ) − 2b2Q2 = 0
with the solutions
∗
Q=
1 ( a1 − α ) =
2b1 , Q2
∗
( a2 − α ) 2b2
 Furthermore, π 11′′ (Q1 , Q2 ) =
−2b1 , π 12′′ (Q1 , Q2 ) = ′′ (Q1 , Q2 ) =
0, and π 22 −2b2 .
 Hence, for all (Q1 , Q2 )
π 11′′ ≤ 0, π 22
′′ ≤ 0, and π 11′′ π 22
′′ − (π 12′′ ) 2 =4b1b2 ≥ 0
then the pair (Q1∗ , Q2∗ ) really does maximize profits.
 The corresponding prices can be found by inserting these values.
P1∗ =a1 − b1Q1∗ =2 ( a1 + α ) , P2∗ =a2 − b2Q2∗ =2 ( a2 + α )
1 1

 Hence, the optimal prices are

P1∗ = 1
2 ( a1 + α ) = 12 (1 + 5) = $3, P2∗ = 1
2 ( a2 + α ) = 12 ( 2 + 5) = $3.5
3. Find the extreme values for f ( x, y ) defined over S when
f ( x, y ) = x 2 + y 2 + y − 1, S= {( x, y) : x 2
+ y 2 ≤ 1} .
(Solution)
 The set S consists of all the points on or inside the circle of radius 1
centred at the origin.
 The continuous function f will attain both a maximum and
a minimum over S , by the extreme value theorem.
 According to the preceding recipe, we start by finding all the stationary points
in the interior of S .
 These stationary points satisfy the two equations
f1′( x, y ) = 2 x = 0, f 2′( x, y ) = 2 y + 1= 0
 So ( x, =
y ) (0, − 1 2) is the only stationary point, and it is in the interior of S ,
with f (0, − 1 2) = − 5 4.
 The boundary of S consists of the circle x 2 + y 2 =
1.

 Note that if ( x, y ) lies on this circle, then in particular both x and y lie
in the interval [−1, 1].
 Inserting x 2 + y 2 = 1 into the expression for f ( x, y ) shows that,
along the boundary of S , the value of f is determined by
the following function of one variable:
g ( y ) = 1 + y − 1 = y, y ∈ [−1, 1]
 The maximum value of= g is 1 for y 1,= and then x 0.
 The minimum value is − 1 when y = −1, and then again x =
0.
 We have now found the only three possible candidates for extreme points,
namely, (0, − 1 2), (0, 1), and (0, − 1).
 But f (0, − 1 2) =− 5 4 , f (0, 1) =
1, and f (0, − 1) =−1.
 We conclude that the maximum value of f in S is 1, which is attained at (0, 1),
whereas the minimum value is − 5 4 , attained at (0, − 1 2).
B. Linear Regression
1. We are given 4 sample data of (demand, price) where xt denotes the price in period t and yt denotes the demand
in period t as shown in the table.

(a) Fill in the blanks below.

number of sample T 4
sum of x Σ xt 8
mean of x µx 2
sum of square of x Σ xt2 18
mean square of deviation, variance of x σxx 0.5

sum of y Σ yt 12
mean of y µy 3
sum of square of y Σ yt2 50
mean square of deviation, variance of y σyy 3.5

sum of product of x and y Σ xtyt 25

covariance of x and y σxy 0.25

slope of best-fit linear function β 0.5

y-intercept of best-fit linear function α 2

(b) What is the linear function that best fits to the date of the table.

Solution
y= 2 + 0.5 x
2. We are going to find a stationary point (α, β) to the function L(α, β):

1
L(α , β= ( 3 − α − 2β ) + ( 4 − α − β ) + ( −α − 2β ) + ( 5 − α − 3β ) 
2 2 2 2
)
4

Applying the regression technique, we seek a linear function that gives best fit to the data.

(a) Find L1′ and L2′ .

Solution
1
 L(α , β= ( 3 − α − 2β ) + ( 4 − α − β ) + ( −α − 2β ) + ( 5 − α − 3β ) 
2 2 2 2
)
4 
∂L 1
 L1′= =  2 ( 3 − α − 2 β ) (−1) + 2 ( 4 − α − β ) (−1) + 2 ( −α − 2 β ) (−1) + 2 ( 5 − α − 3β ) (−1) 
∂α 4 
∂L 1
 L2′ = =  2 ( 3 − α − 2 β ) (−2) + 2 ( 4 − α − β ) (−1) + 2 ( −α − 2 β ) (−2) + 2 ( 5 − α − 3β ) (−3) 
∂β 4 

(b) Find a stationary point.

Solution
1
  2 ( 3 − α − 2 β ) (−1) + 2 ( 4 − α − β ) (−1) + 2 ( −α − 2 β ) (−1) + 2 ( 5 − α − 3β ) (−1)  =0
4
1
  2 ( 3 − α − 2 β ) (−2) + 2 ( 4 − α − β ) (−1) + 2 ( −α − 2 β ) (−2) + 2 ( 5 − α − 3β ) (−3)  =0
4

 2α − 6 + 4 β =0
 8β − 12 + 4α + β − 0.5 =
0
⇒= β 0.5, = α 2

 We have a stationary point (α , β ) = (2, 0.5).

C. Constrained Optimization: Lagrange Multiplier Method
1. max(min) 3xy
subject to
x2 + y 2 = 8
(Solution)
( )
 With  ( x, y ) = 3 xy − λ x 2 + y 2 − 8 , the first-order conditions are
1′( x, y ) =3 y − 2λ x =0 and
2′ ( x, y ) =3 x − 2λ y =0.
=
 These can be rewritten as (i) 3 y 2=λ x and (ii) 3 x 2λ y.
 If x 0,=then (i) gives y 0;= conversely, if y 0, then (ii) gives x = 0.
 But ( x, y ) = ( 0, 0 ) does not satisfy the constraint.
 Hence x ≠ 0 and y ≠ 0.
 Equating the ratio of the left-hand sides of (i) and (ii) to the ratio of their right-hand sides,
one has y x = x y or x 2 = y 2 .
 Finally, using the constraint gives x=
2
y=
2
4.
 The four solution candidates are therefore
( 2, 2 ) and ( −2, −2 ) with λ =3 2,
( 2, −2 ) and ( −2, 2 ) with λ = −3 2.
 The corresponding function values are
f ( 2, 2 ) = f ( −2, −2 ) = 12 and
f ( 2, −2 ) =f ( −2, 2 ) =−12.
 We have minimum at (2, − 2), (−2, 2) and maximum at (2, 2) and (−2, −2).

D. From a student’s question

1. Show that
ln 2 − ln(e −1 3 + 1) = 13 + ln 2 − ln(e1 3 + 1).

Solution.
 We need to show that 1
3 − ln(e1 3 + 1) =− ln(e −1 3 + 1)
 1
3 − ln(e1 3 + 1)
= ln(e1 3 ) − ln(e1 3 + 1)
=− ln(e1 3 + 1) − ln(e1 3 ) 
e1 3 + 1
= − ln
e1 3
= ln(1 + e −1 3 )