Help with production optimization problem?

Rumor

New member
Joined
Sep 8, 2009
Messages
27
So, here's the full problem:

"The production amount, Q, of an item manufactured by a company is modeled by the CobbDouglas function

Q = 200(K^0.6)(L^0.4)(T^0.3)

where K is the quantity of capital investment, L is the quantity of labor used, and T is the investment in training. Capital costs are $17 per unit, labor costs are $24 per unit, and training costs are $8 per unit. The company wants to keep the total cost at the current $20,000 budget while producing as much of the item as possible. Suppose you are asked to consult for the company, and you learn that 800 units of capital, 200 units of labor, and 200 units of training are being used.

What do you advise? Should the plant use more or less labor? More or less capital? More or less training? By what percent can the company increase production over the current amount without changing the budget? At the optimal allocation, by approximately how much will the production quantity change with a dollar increase in the total budget? Answer this without computing a new solution, but do give an explanation. At the optimum allocation, what is the marginal cost of producing an item (that is, the cost of an additional item)?"

From what I can gather, the constraint here is 17K+24L+8T=2000, so the resulting equation would be Q = 200(K^0.6)(L^0.4)(T^0.3) - (17K+24L+8T). Our professor encourages us to use a program called Maple 10, and by using that, I've gotten the partial derivatives with respect to x, y, z.

Here are the program's results:

fx = ((120.0 * L^0.4 * T^0.3)/(K^0.4)) - 17
fy = ((80.0 K^0.6 T^0.3)/(L^0.6))-24
fz = ((60.0 K^0.6 L^0.4)/(T^0.7))-8

My next step was to solve these equations by setting them to 0. But after that, I'm not sure what to do, nor am I certain that what's been done so far is correct. Any input would be appreciated!
 
Rumor said:
So, here's the full problem:

"The production amount, Q, of an item manufactured by a company is modeled by the CobbDouglas function

Q = 200(K^0.6)(L^0.4)(T^0.3)

where K is the quantity of capital investment, L is the quantity of labor used, and T is the investment in training. Capital costs are $17 per unit, labor costs are $24 per unit, and training costs are $8 per unit. The company wants to keep the total cost at the current $20,000 budget while producing as much of the item as possible. Suppose you are asked to consult for the company, and you learn that 800 units of capital, 200 units of labor, and 200 units of training are being used.

What do you advise? Should the plant use more or less labor? More or less capital? More or less training? By what percent can the company increase production over the current amount without changing the budget? At the optimal allocation, by approximately how much will the production quantity change with a dollar increase in the total budget? Answer this without computing a new solution, but do give an explanation. At the optimum allocation, what is the marginal cost of producing an item (that is, the cost of an additional item)?"

From what I can gather, the constraint here is 17K+24L+8T=2000, so the resulting equation would be Q = 200(K^0.6)(L^0.4)(T^0.3) - (17K+24L+8T). Our professor encourages us to use a program called Maple 10, and by using that, I've gotten the partial derivatives with respect to x, y, z.

Here are the program's results:

fx = ((120.0 * L^0.4 * T^0.3)/(K^0.4)) - 17
fy = ((80.0 K^0.6 T^0.3)/(L^0.6))-24
fz = ((60.0 K^0.6 L^0.4)/(T^0.7))-8

My next step was to solve these equations by setting them to 0. But after that, I'm not sure what to do, nor am I certain that what's been done so far is correct. Any input would be appreciated!

You are close to the right path, but NOT on it.

I have not worked with Cobb-Douglas production functions in a very long time so I am quite rusty. We shall have to work this through together, OK?

From what I can gather, the constraint here is 17K+24L+8T=2000 No, the budget is 20,000 so the constraint is 17K+24L+8T=20,000. Was this a typo?

You don't quite have the hang of Lagrangian multipliers. You set up a new function P = Q + W(C), where C is the constraint function. You maximize P to get the maximum of Q SUBJECT to the constraint. The W is your Langrangian multiplier. When you calculate the partial with respect to W, you get C as one of the conditions of the maximum, which is what you want. Do you understand. So, Q = 200(K^0.6)(L^0.4)(T^0.3) - (17K+24L+8T) is not quite right because you are missing the multiplier W (it is usually specified as the Greek letter lambda, but I have no clue on how to get the Greek alphabet on this site so I'll use W.)

Furthermore, you may not understand the purpose of the multiplier. When you take the partial of P with respect to W and equate the partial to 0, the W disappears, leaving behind the constraint function. So the constraint function has to be [20,000 - (17K + 24L + 8T)]. When you you differentiate
W[20,000 - (17K + 24L + 8T)] with respect to W you get [20,000 - (17K + 24L + 8T)]. When you equate that to zero, you get 20,000 = 17K + 24L + 8T. Your constraint is now part of the maximization process. With me?
 
Top