Empirical Industrial Organization Problem Set 2

Dynamic Discrete Choice

In this assignment you have to simulate a simple version of an inventory control problem. The objective is to familiarize with the computation of the dynamic programming solution and compare it with a CCP method.

Deadline: Please send your problem set before January 10th, 2025.

Requirements

see here

Remarks

There are two types of penalty term in the utility function, each with their economic interpretation.

Case 1: $-\lambda 1{c>0} 1{i=0}$

1. In this period t, I somehow want to eat chips $c>0$, but I have run out of chips $i=0$.
2. I incur some penalty because I can not eat my fav snack right away. I need to go out and buy some.
3. Then I go to the supermarket, see the price $(p_s)$. I know the state $(i=0,c>0,p, \epsilon)$ when making the decision $x$.
4. If I decide to buy $x=1$, then my next period inventory would be $i'=i+x-c=0.75$
5. If I decide not to buy $x=0$, then my next period $i'=\max{0,i+x-c}=0$

Case 2: $-\lambda 1{c>0} 1{i+x=0}$

1. In this period t, I somehow want to eat chips $c>0$, but I have run out of chips $i=0$.
2. Then I go to the supermarket, see the price $(p_s)$. I know the state $(i=0,c>0,p, \epsilon)$ when making the decision $x$.
3. If I decide to buy $x=1$, then I can happily eat the chips and there's no penalty. My next period $i'=i+x-c=0.75$
4. If I decide not to buy $x=0$, I incur the loss of "starving". Then my next period $i'=\max{0,i+x-c}=0$

Code

see here

Report

see here