Successful heuristic algorithms for solving combinatorial optimization problems have mimicked processes observed in nature. Two highly successful families of algorithms that do this are simulated annealing and genetic algorithms. Here, a third family of algorithms, ant colony optimization is explored and implemented in C#. The test bed for evaluating the quality of solutions is based on several Traveling Salesperson Problems (TSP) of varying size using real-world data (Euclidian problems) with known solutions.
Note: Due to the typographical sophistication of this article, no HTML version is available. Please use the PDF version.
About the author
Richard Wiener, "Ant Colony System Optimization", in Journal of Object Technology, vol. 8 no. 6, September-October 2009, pp. 39-58 http://www.jot.fm/issues/issue_2009_09/column4/