Cellular Ant Algorithm and Its Convergence

朱刚,马良,高岩
DOI: https://doi.org/10.3969/j.issn.1671-1815.2009.05.005
IF: 2.034
2008-01-01
Systems Engineering
Abstract:Ant algorithm has successfully solved series of discrete optimization problems.However its global convergence is not fully studied and proved.Cellular Ant Algorithm is a new method for optimization based on the principle of cellular automata.Through redistributing on the pheromone of the ants,the algorithm is improved on the search performance in solution space.This paper gives a convergence analysis for Cellular Ant Algorithm by using the theory of Markov Chains.It is shown that under certain conditions,the solutions generated in each iteration of this graph-based cellular ant system converge with a probability that can be made arbitrarily close to 1.
What problem does this paper attempt to address?