Immune clonal selection algorithm for capacitated arc routing problem

In existing metaheuristics for solving the capacitated arc routing problem, traversal local search operators are often used to explore neighbors of the current solutions. This mechanism is beneficial for finding high-quality solutions; however, it entails a large number of function evaluations, causing high computational complexity. Hence, there is a need to further enhance the efficiency of such algorithms. This paper proposes a high-efficiency immune clonal selection algorithm for capacitated arc routing instances within a limited number of function evaluations. First, an improved constructive heuristic is used to initialize the antibody population. The initial antibodies generated by this heuristic help accelerate the algorithm’s convergence. Second, we show how an immune clonal selection algorithm can select in favor of these high-quality antibodies. By adopting a variety of different strategies for different clones of the same antibody, it not only promotes cooperation and information exchanging among antibodies, but also increases diversity Communicated by V. Loia. R. Shang (B) · H. Ma · J. Wang · L. Jiao Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education of China, Xidian University, Xi’an 710071, China e-mail: [email protected] L. Jiao e-mail: [email protected] J. Wang Xi’an Institute of Huawei Technology Co. Ltd., Xi’an 710075, China e-mail: [email protected] R. Stolkin School of Computer Science, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK e-mail: [email protected] and speeds up convergence. Third, two different antibody repair operations are proposed for repairing various kinds of infeasible solutions. These operations cause infeasible solutions to move towards global optima. Experimental studies demonstrate improved performance over state-of-art algorithms, especially on medium-scale instances.