Discrete symbiotic organism search with excellence coefficients and self-escape for traveling salesman problem

Abstract

Traveling salesman problem (TSP) is one of the well-known NP-hard problems in combinatorial optimization. The optimal solution of large-scale TSP is difficult to find with exact algorithms in acceptable computation time. Thus, heuristic algorithms are widely studied to find the near-optimal or satisfactory solutions. The discrete symbiotic organism search (DSOS) is enhanced with excellence coefficients and self-escape strategy (ECSDSOS) for TSP. The excellence coefficients let ECSDSOS choose shorter edges (routes) for generating better local paths. The local optimization capability is improved so that ECSDSOS is accelerated for finding satisfactory solutions. The self-escape strategy improves the organism versatility in order to prevent ECSDSOS from being trapped into local minima. The premature convergence is constrained and the global optimization capability is improved. ECSDSOS is tested with certain real-world TSP instances in TSPLIB. The results are compared with those computed with basic DSOS and several other heuristic algorithms. Under the same preconditions, ECSDSOS shows competitive optimization capacities for computing better solutions for TSP than basic DSOS and some other heuristic algorithms.