Ant algorithms are a recently developed, population-based approach which has been successfully applied to several NP-hard combinatorial optimization problems. In this paper, through an analysis of the constructive procedure of the solution in the Ant Colony System (ACS),we present an Ant Colony System Hybridized with Randomized Algorithm( RAACS). In RAACS, only partial cities are randomly chosen to compute the state transition probability. Experimental results for solving the Traveling Salesman Problems( TSP) with both ACS and RAACS demonstrate that averagely speaking, the proposed method is better in both the quality of solutions and the speed of convergence compared with the ACS.
Index Terms:
Ant Colony System; Combinatorial Optimization; Randomized Algorithm; TSP
Citation:
Chengming Qi, "An Ant Colony System Hybridized with Randomized Algorithm for TSP," snpd, vol. 3, pp.461-465, Eighth ACIS International Conference on Software Engineering, Artificial Intelligence, Networking, and Parallel/Distributed Computing (SNPD 2007), 2007