Affiliations: [a] Intelligent Information Processing Research Center, Fujian University of Technology, Fuzhou, Fujian, China
| [b] Fujian Provincial Key Laboratory of Big Data Mining and Applications, Fujian University of Technology, Fujian, China
| [c] College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, China
| [d] Department of Information Technology, University of Manage and Technology, Haiphong, Vietnam
Abstract: Metaheuristic algorithms have been applied widely for real-world problems in many fields, e.g., engineering, financial, healthcare. Bats algorithm (BA) is a recent metaheuristic algorithm with considering as a robust optimization method that can outperform existing algorithms. However, when dealing with complicated combinatorial problems such as traveling salesman problems (TSP), the BA can be fallen in a local optimum. This paper proposes a new hybridizing Parallel BA (HPBA) with a mutation in local-search to escape such its drawback scenario for TSP. A graph theory mutation method is used to embed for hybridizing BA with exploiting similarities among individuals. The proposed method is extensively evaluated in TSP with series instances of the benchmark from TSPLIB to test its performance. The compared experimental result with the previous method and the best-known solutions (B.K.S) in the literature shows that the proposed approach offers competitive results.