Rank-based memetic algorithm for capacitated arc routing problems

The capacitated arc routing problem (CARP) has garnered much attention recently, owing to its wide range of applications in the real world. This study describes an efficient memetic algorithm for solving CARP. First, the concepts of Rank Number (RankNo), and Rank Count (RankCount) are proposed, to shed light onto the edge selection rules. Then, the essential backbone of the algorithm, the Rank-based Neighborhood Search (RENS) operator, is introduced. Based on these above-mentioned concepts, methods relating to selection and evaluation of edge(s) are designed to make local search more effective. Two rules, namely mapping rule (MAR) and move rule (MOR), are constructed explain the working of the RENS operator. Finally, this algorithm is tested on seven famous benchmark sets. The experimental results show that it has the better performance than two compared state-of-the-art algorithms.