Vehicle Scheduling Optimization Based On Hybrid Genetic Algorithm

Genetic algorithms are a class important global stochastic optimization methods.This algorithm is often used to solve vehicle scheduling optimization problems because it does not require continuation and derivative, and are easy to operate.The main method of this dissertation is to take the fixed crossover and mutation rate adjustment strategies of the traditional genetic algorithm with an adaptive strategy to give an improved genetic algorithm replaced, and combined with the local search capacity of tabu search algorithm is strong features, an improved hybrid genetic algorithm is presented. Through in-depth research on vehicle scheduling problem, multi-depot VSP problem with constraints converted to its equivalent cycling field VSP problem with constraints. The hybrid genetic algorithm is studied model of cycling field VSP and model of multi-depot field VSP. And in time windows of VSP, at first, through the FSP algorithm to find a good initial solution, reuse TS algorithm local search capability optimal or satisfactory solution, and adaptive GA algorithm global optimization ability to get the question, finally, validate the hybrid GA algorithm by numerical results.The main contributions are listed as follows:In Chapter 1, the research status of some well-known of solving vehicle scheduling optimization problem methods are reviewed.In Chapter 2, inspired by the articles of Lang Maoxiang and Jian Jie and Liu Wusheng etc, presents a cross with adaptive mutation rate instead of the traditional genetic algorithm with fixed crossover and mutation rate improved genetic algorithm approach a hybrid genetic algorithm, combined with the TS algorithm to solve cycling field VSP model, and by numerical results demonstrate the e?ectiveness of this method.In Chapter 3, an improved hybrid genetic algorithms of solving cycling field vehicle scheduling optimization problem is presented on the basis of a hybrid genetic algorithm given by Wang Xiaobo, proposed a hybrid coding, combined with the TS algorithm and climbing algorithm improved hybrid genetic algorithm for multi-depot VSP model.Finally, numerical results show that the algorithm is e?ective.In Chapter 4, a hybrid genetic algorithm for solving Multi-Depot vehicle scheduling optimization problem with time window are presented noting by the articles of Zhong Shiquan etc and Liu Jiali etc. the first by FSP algorithm to find a good initial solution to improve the local search ability TS algorithm, secondly, proposed a yard, vehicles and customers using natural numbers coding, and combined with scanning algorithm and CW algorithm gives an initial population structure improved hybrid genetic algorithm for solving multi-depot VSP model with time window, Finally, by numerical results show the e?ectiveness of the algorithm.