Implementing Dynamic Programming

Abstract

We got acquainted with the DP decomposition principle and Bellman’s recursive equation in Chap. 1. Now we turn to practical matters in terms of actual implementation and computational issues. As we have stressed, dynamic programming is a principle, rather than a specific algorithm that can be purchased off-the-shelves. Thus, careful modeling and analysis are needed on the one hand, and customized implementation is needed on the other one. Indeed, one of the main aims of this tutorial booklet is to encourage newcomers to experiment with DP, which is only possible by implementing it using any programming language or computing environment they like. Our choice is MATLAB, which is widely available and quite convenient as it yields short and intuitive code, not obscuring the underlying ideas.