Differential Ant-Stigmergy Algorithm for Optimal Control Problems

Abstract

The objective of this paper is to show how the Differential Ant-Stigmergy Algorithm (DASA) can be used to solve an Optimal Control Problem (OCP). A version of this algorithm devoted to search a good solution for an OCP is presented. Because this kind of problems has usually a large computational complexity, a technique to reduce it is presented. This one exploits a particularity of DASA related to the coding of the problem's solution. In the first phase, DASA find out a good solution in a small number of iteration by adopting a rough coding of the control input. The second phase is a kind of zoom of the search space around this solution, which becomes initial solution, with a finer coding of the control input.