Research On Improved Glowworm Swarm Optimiza-Tion In UAV Vehicle Joint Delivery Path Problem

In recent years,the application of unmanned aerial vehicles(UAVs)in delivery has emerged as an efficient,safe,and environmentally-friendly alternative to traditional vehicle-based delivery methods.UAVs offer advantages such as high speed,low cost,and minimal operational requirements,enabling rapid delivery.However,they are limited in terms of payload capacity and endurance.On the other hand,vehicles are capable of transporting large quantities of goods over long distances but are constrained by traffic conditions and road quality.As a new delivery approach,the integration of UAVs and vehicles in a joint delivery system allows both methods to leverage their respective strengths,resulting in fast,efficient logistics and emergency response.The optimization of UAV-vehicle joint delivery routes has emerged as a prominent research topic within the field of logistics in recent years.This paper focuses on investigating the path problem within a UAV-vehicle joint delivery system,where a single vehicle is assigned to a single UAV for synchronized delivery.To address this problem,a Mixed Integer Linear Programming model(MILP)is employed to accurately describe the optimization task.Furthermore,a customized Glowworm Swarm Optimization(GSO)specifically tailored for the UAV-vehicle routing problem is devised to optimize the combined paths of UAVs and vehicles.This research endeavor encompasses the following key research contributions:(1)Environment Modeling and Dataset Construction.In the realm of map environment modeling,a systematic approach is employed to tackle both known and unknown road conditions for effective UAV-vehicle joint delivery route planning.For regions with readily available road condition data,the Gaode map service is leveraged to acquire essential information concerning road conditions and environmental factors between customer nodes.In instances where road conditions are ambiguous,an innovative solution involving drone aerial photography is implemented.Through the utilization of drones,high-resolution images are captured to capture detailed road information.The collected data is then processed using a grid-based methodology to construct an accurate map representation of the region of interest.Subsequently,the A*algorithm is employed to derive the optimal path information between customer nodes.This comprehensive approach ensures that all customer nodes are adequately connected,enabling the successful construction of a robust dataset for the UAV-vehicle joint delivery route problem.(2)The MILP Model for the UAV-vehicle j oint delivery problem path optimization.Through the analysis of the UAV-vehicle joint delivery route problem of single-vehicle and single-UAV synchronous delivery,considering the complexity of the UAV-vehicle joint delivery problem,a hypothesis is put forward for the problem and the model parameters are set in combination with the actual situation.Based on the distribution cost analysis,a multi-objective optimization function about time cost and energy cost is established.For the traditional UAV single-point delivery,the paper proposes that the UAV can deliver multiple nodes at a time to improve delivery,and at the same time design the corresponding constraints.(3)Discrete Glowworm Swarm Optimization(DGSO)design for the UAV-vehicle routing problem.In order to solve the routing problem of joint distribution of UAVs and vehicles,GSO,which was originally suitable for continuous domains,was improved by discretization.Among them,an integer sequence code is used to describe individual fireflies,and the Hamming distance between two individuals is used to quantify the difference between them.New search operators and moving methods are designed to generate new firefly individuals to update the population.Among them,in order to avoid the algorithm from falling into local optimum,the position of firefly is perturbed.(4)Simulation.The correctness of the above research methods is verified by simulation experiments.Aiming at the two methods of UAV multi-node combined delivery and traditional single-point delivery,the DGSO designed in this paper is used to carry out cost calculation and comparative analysis of the UAV-vehicle j oint delivery path problem of example a801 and a1002.The simulation results prove that the distribution cost is lower by adopting the UAV multi-node combination distribution method.The Non-dominated Sorting Genetic Algorithm(NSGA)and the DGSO designed in this paper are used to solve different examples.The result proves that the performance of the DGSO designed in this paper is better than that of the NSGA.Finally,the paper analyzes the sensitivity factors of the problem,the influence of the UAV’s speed,load capacity,and endurance time on the UAV-vehicle joint delivery problem.