Research On Muti-Objective AGV Routing Optimization Considering Battery Recharge And Conflict-Free Constraints

With the increase of global trade volume and maritime demand,the development of Automated Container Terminals is the current trend of global terminals.Terminal operators begin to pay more attention to improving the operating efficiency of terminals and reducing the operating costs of terminals.Therefore,the reasonable scheduling and planning of AGV driving routes for container terminals is the optimal direction to help automated container terminals improve operational efficiency.Firstly,this paper describes the development status of ACT and analyzes the container operation of AGV at the terminal in detail.Combining the VRP problem with the AGV scheduling of automated container terminals,taking the power limit and conflict-free route planning into account,it studies the scheduling problem of AGV at multi-objective ports.Considering that the automatic terminal AGV is driven by electricity,the power limit of the AGV needs to be further analyzed: when the AGV battery power is lower than the safety threshold,it is necessary to go to the power exchange station to ensure the power exchange operation.Different from the traditional VRP problem,the driving range of AGV is a fixed road network.As the number of tasks and AGV operations increase,there will be potential conflict risks.Based on the above analysis,a multi-objective optimization model for port AGV scheduling considering power exchange and non-conflict constraints is established.The optimization objective is to achieve the maximum customer satisfaction while the maximum completion time is the shortest.Then,a two-stage algorithm for solving the above multi-objective model is designed.In the first stage,the multi-objective particle swarm optimization algorithm is used to schedule and assign tasks between tasks and AGVs,and the optimization is carried out with the shortest completion time and the maximum customer satisfaction as the two objectives.In the second stage of planning,Dijkstra algorithm is used to obtain the shortest distance for AGV to complete the task,and based on the way of prediction time window,conflict resolution strategy is adopted to dynamically adjust the time window of the nodes that AGV has traveled to avoid the conflict between AGVs.With the increase of the number of AGVs and tasks,it is necessary to further calculate the time of power exchange and adjust the time of AGV to the power exchange station to reduce the waiting time of AGV in the power exchange station.Finally,the multi-objective particle swarm optimization algorithm,multi-objective genetic algorithm and improved particle swarm optimization algorithm are realized by programming with MATLAB,and ten examples of three scales are designed for simulation testing.The analysis is made from the average value of the solution set,Pareto frontier distribution and multi-objective indicators(HV and IGD).The test results show that the performance and results of the improved particle swarm optimization algorithm are better than the other two algorithms in different scales.Port operators can use this algorithm to get better scheduling strategy.