A Discrete Monarch Butterfly Optimization for Chinese TSP Problem

Recently, Wang et al. proposed a new kind of metaheuristic algorithm, called Monarch Butterfly Optimization (MBO), for global continuous optimization tasks. It has experimentally proven that it has better performance than some other heuristic search strategies. On the other hand, travelling salesman problem (TSP) is one of the most representative NP-hard problems that are hard to be solved by traditional methods. It has been widely studied and solved by several metaheuristic algorithms. In this paper, MBO is discretized, and then a discrete MBO (DMBO), and firstly used to solve Chinese TSP (CTSP). In the basic MBO, Wang et al. had made little effort to fine-tune the parameters. In our present work, the parametric study for one of the most parameter, butterfly adjusting rate (BAR), is also provided. The best-selected BAR is inserted into the DMBO method and then solve CTSP problem. By comparing with three other algorithms, experimental results presented clearly demonstrates DMBO as an attractive addition to the portfolio of swarm intelligence techniques.