This paper describes an efficient hybrid genetic algorithm (HGA) for the tr
aveling salesman problem. In general, a genetic algorithm (GA) combined wit
h other algorithms (e.g., a local search) is well known to be a powerful ap
proach. The other algorithms are divided into local search heuristics and m
etaheuristics. In incorporating the metaheuristics, it is reported that a d
ifficult problem of changing processes between the GA process and the metah
euristic search process appears. To avoid the difficult problem using simul
ated annealing as one of the metaheuristics, we investigate an efficient HG
A that does not involve the problem. (C) 2000 Scripta Technica, Electron Co
mm Jpn Pt 3, 84(2): 76-83, 2001.