Ant Colonies Discover Knight's Tours

摘要

In this paper we introduce an Ant Colony Optimisation (ACO) algorithm to find solutions for the well-known Knight's Tour problem. The algorithm utilizes the implicit parallelism of ACO's to simultaneously search for tours starting from all positions on the chessboard. We compare the new algorithm to a recently reported genetic algorithm, and to a depth-first backtracking search using Warnsdorff's heuristic. The new algorithm is superior in terms of search bias and also in terms of the rate of finding solutions.