GA applied method for interactively optimizing a large-scale distribution network

Based on experimental comparison, this paper discusses approximate solution methods of the medium-scale traveling salesman problem (TSP) that suit repetitive use in interactive simulation for globally optimizing a large-scale distribution network. For example, such a nationwide distribution network consists of approximate 1000 trucks. So, the optimization of such a distribution network needs repetitive interactive simulations, in each of which about 1000 TSPs are automatically solved after changing such conditions as the area division and the truck allocation and a human user waits and checks the results totally. Therefore, our proposed method applied genetic algorithms (GA), guarantees interactive responsiveness and realizes experts' level accuracy, through enabling to solve 1000 middle scale TSPs for a distribution network within 30 seconds within 3% errors. Experimental results proved that the proposed method enables to optimize a large-scale distribution network.