Abstract:

On the basis of establishing an optimized model for the physical distribution routing problem, the paper uses techniques such as maximum preserved rou lette wheel selection, maximum preserved crossover and self adaptation for probability of crossover and mutation, designs a genetic algorithm based on n  atural numbers, and performs experimental calculations. The results demonstrate that the feasible solutions to the not-fully-loaded vehicle scheduling p roblem with soft time windows can easily be obtained using the algorithm.

Key words: （soft time window, physical distribution, routing selection, genetic algorithm）