Hybrid evolutionary algorithms for the Multiobjective Traveling Salesman Problem

摘要: The multiobjective TSP is solved using three hybrid evolutionary algorithms.Two of them are based on Differential Evolution, the third one NSGA II.All algorithms used in a hybridized form with VNS algorithm.The solve 2-5 objective functions TSP.Comparisons between all were performed set benchmark instances. In Multiobjective Traveling Salesman Problem (moTSP) simultaneous optimization more than required. this paper, common characteristics proposed and analyzed for solution Problem. two Evolution algorithm version II. One challenges efficient application an algorithm, which suitable continuous problems, combinatorial problem. Thus, we test different versions use external archive (denoted as MODE) other crowding distance NSDE). Also, another novelty mutation operators each leading to six (MODE1, MODE2 MODE3 first NSDE1, NSDE2 NSDE3 second version). We Variable Neighborhood Search (VNS) procedure separately order increase exploitation abilities algorithms. give quality algorithms, experiments conducted classic Euclidean instances taken from library. number evaluation measures conclude most selected general, can easily be applied routing problems by changing function constraints problem they have ability (in paper up five functions). global search increases exploration giving very effective