A new coding method for genetic algorithm in vehicle routing problem

Abstract

Genetic algorithm is widely used to solve optimization problems for its characteristic, especially vehicle routing problem. By using this algorithm to solve the vehicle scheduling problem, determine the solution coding method is a very crucial basic work, because it will directly influence the algorithm to realize and the inferiority of algorithm performance. This paper presents a new coding method — the client direct arrangement and grouping randomized. Firstly, produce customer arrangement without repeating, and than put it into small groups randomly which is serviced by a vehicle separately. It represents a solution or a route. In the final, it is proved that the genetic algorithm with new coding has good performance through experiment.