Performance analyses over population seeding techniques of the permutation-coded genetic algorithm

摘要: A review on population seeding techniques for Permutation-coded Genetic Algorithm.Experiments are performed large sized Traveling Salesman Problem instances.The experimentation analyses carried out using ANOVA and DMRT statistical tools.Best performing identified w.r.t. assessment criteria. The genetic algorithm (GA) is a based meta-heuristic global optimization technique dealing with complex problems very search space. initialization crucial task in GA because it plays vital role the convergence speed, problem space exploration also quality of final optimal solution. Though importance deciding specific widely recognized, hardly addressed literature. In this paper, different permutation-coded such as random, nearest neighbor (NN), gene bank (GB), sorted (SP), selective (SI) along three newly proposed ordered distance vector have been extensively studied. ability each has examined terms set performance criteria computation time, rate, error average convergence, diversity, nearest-neighbor ratio, distinct solutions distribution individuals. One famous combinatorial hard traveling salesman (TSP) being chosen testbed experiments benchmark TSP instances obtained from standard TSPLIB. tools to claim unique characteristic best defined nature application.