An improved hybrid algorithm for traveling salesman problem

Abstract

An improved artificial bee colony (ABC&VNS) to settle the traveling salesman problem (TSP) is described in this paper. This hybrid algorithm's principal character is to incorporate the solution construction mechanism of the artificial bee colony (ABC) with variable neighborhood search (VNS) to exploit the intrinsic structure of the solutions to acquire the high quality solution. Besides, three novel neighborhood structures are present, in which problem-specific knowledge is utilized respectively, to improve both the search efficiency and quality of VNS. Finally, the experiments for benchmark TSP instances are carried out to verify that the proposed algorithm is capable of maintaining high performance to solve the TSP, which is compared with the best existing approaches in the way of solution quality.