6. THE HOWARD POLICY ITERATION ALGORITHM
In discrete event systems, a graph or petri net is often used to show the dependence of various events. Such events could be the production of a "unit" in a factory production line, or the departure of a car on a road trip. Discrete event systems generally correspond loosely to "man-made" systems, as opposed to natural systems which seem to be based more on the mathematics of differential equations and chaos theory. In such a system, IF its operation eventually becomes periodic (as turns out in most cases), the Howard Policy Iteration Algorithm calculates this cycle time. It is also an algorithm used in the general context of solving Markov Decision Processes (see http://victoria.mindmaker.hu/~szepes/research/Summary.htm ). This project would focus on an understanding of the algorithm, its implementation in some programming language, its application to some sample cases and a comparison to other algorithms that tackle this problem. Knowledge of/interest in programming and/or mathematics would benefit in this work.