Research On Multi-vehicle Dispatching Method Based On Improved Genetic Algorithm

The rapid development of unmanned systems makes their application scenarios more extensive.Unmanned vehicles are a type of unmanned systems.Due to the limited carrying capacity of a single unmanned vehicle,a multi-vehicle system is generally used for multi-tasks.This brings up the problem of how to perform task schedule so that the system can achieve the lowest transportation cost under the premise of safety.That is,the problem of multi-vehicle dispatching in unmanned systems.Meanwhile,due to different application scenarios,dispatching problems will have different specific constraints.At present,the existing multi-vehicle dispatching algorithms still has bottlenecks.For the scenarios where multiple orders are scheduled at the same time,in the three dimensions of the system’s total dispatching routes’ length,total dispatching time and the time of completing missions,the explainable system optimization has not yet reached.There is still room for optimization.At the same time,for typical application scenarios such as the AGV dispatching system with path constraints,it is necessary to generate an optimal dispatching solution with conflict-free routes.For unmanned delivery vehicle dispatching system with time constraints,it is necessary to generate optimized dispatching routes with time windows.Therefore,this paper studies the dispatching algorithm of the multi-vehicle system,proposes an improved genetic algorithm for multi-vehicle dispatching.At the same time,it analyzes the AGV scheduling system and the unmanned delivery vehicle dispatching system in the multi-vehicle system.This paper proposes an algorithm to optimize the corresponding systems under different constraints,and realizes its application in these two systems.First,this paper investigates and classifies the existing multi-vehicle dispatching algorithms,analyzes the genetic algorithm and ant colony algorithm.And optimizes the two algorithms according to their advantages and disadvantages.An adaptive genetic algorithm evolution factor and ant colony algorithm pheromone update method are presented.In order to improve the efficiency of the algorithm,an improved genetic algorithm that fuses the genetic algorithm with the ant colony algorithm and an adaptive algorithm switching factor and pheromone concentration initialization calculation method are designed.The two algorithms are cascaded in a certain way.Experiments on standard data sets prove that the improved genetic-ant-colony algorithm has better computing performance and practicability.Subsequently,by analyzing different multi-vehicle systems,the AGV dispatching system and the unmanned express delivery dispatching system are selected for analysis.The proposed algorithm was optimized in different directions in these two systems.For the advancement of the AGV dispatching system,a path optimization method based on time-space graph search is proposed.Based on the improved genetic algorithm for dispatching,the AGVs’ paths are optimized to avoid path conflicts.For the unmanned express delivery vehicle system,an improved genetic algorithm based on the time window is used,which takes the time cost as one of the factors for the fitness calculated by the algorithm.The capacity limit is quantified as a conditional constraint,so that the function meets the capacity limit in the calculation process.The simulations prove the feasibility and efficiency of the proposed improved genetic algorithm and its optimization method.