A Real Adjacency Matrix-Coded Differential Evolution Algorithm for Traveling Salesman Problems

Permutation-based combinatorial optimization problems have very wide application. Aim of the study is to design a real coding mechanism for evolutionary computing to solve permutation-based COPs. A real adjacency matrix-coded differential evolutionary algorithm (RAMDE) is proposed to solve traveling salesman problem (TSP): a classic COP. Considering TSP structure, a swarm of real adjacency matrices is adopted to represent individuals within population and arithmetical operators of DE execute in form of real matrices. Experimental results show that the proposed real adjacency matrix-coding mechanism is promising to extend DE for COPs.