Knapsack problem

The Knapsack Problem

Let's apply what we're learned so far to a slightly more interesting problem. You are an art thief who has found a way to break into the impressionist wing at the Art Institute of Chicago. Obviously you can't take everything. In particular, you're constrained to take only what your knapsack can hold — let's say it can only hold W pounds. You also know the market value for each painting. Given that you can only carry W pounds what paintings should you steal in order to maximize your profit?

First let's see how this problem exhibits both overlapping subproblems and optimal substructure. Say there are n paintings with weights w₁ , ..., w_n and market values v₁ , ..., v_n . Define A(i,j) as the maximum value that can be attained from considering only the first i items weighting at most j pounds as follows.

Obviously A(0,j) = 0 and A(i,0) = 0 for any i ≤ n and j ≤ W. If w_i > j then A(i,j) = A(i-1, j) since we cannot include the i^th item. If, however, w_i ≤ j then A(i,j) then we have a choice: include the i^th item or not. If we do not include it then the value will be A(i-1, j). If we do include it, however, the value will be v_i + A(i-1, j - w_i ). Which choice should we make? Well, whichever is larger, i.e., the maximum of the two.

Expressed formally we have the following recursive definition

This problem exhibits both overlapping subproblems and optimal substructure and is therefore a good candidate for dynamic programming. The subproblems overlap because at any stage (i,j) we might need to calculate A(k,l) for several k < i and l < j. We have optimal substructure since at any point we only need information about the choices we have already made.

 

http://20bits.com/article/introduction-to-dynamic-programming