Skip to Main Content
This paper presents an agent-based evolutionary search algorithm (AES) for solving dynamic travelling salesman problem (DTSP). The proposed algorithm uses the principal of collaborative endeavor learning mechanism in which all the agents within the current population co-evolve to track dynamic optima. Moreover, a local updating rule which is much the same of permutation enforcement learning scheme is induced for diversity maintaining in dynamic environments. The developed search algorithm and benchmark generator are then built to test the evolutionary model for dynamic versions of travelling salesman problem. Experimental results demonstrate that the proposed method is effective on dynamic problems and have a great potential for other dynamic combinatorial optimization problems as well.