Study On Route Optimization Of Electric Vehicle Sharing Service In Autonomous Driving Environment

In response to the side effects of the transportation industry on energy and the environment,countries around the world are committed to the development of new-energy vehicles.And battery electric vehicles have become the dominant development direction of new energy transportation technology due to their “low noise,no direct exhaust emissions” and other characteristics.At the same time,to alleviate the traffic congestion,many cities introduced shared electric vehicles as a new mode of transportation to relieve the tension on road resources.However,problems such as the inconvenience of finding a car or parking,and the difficulty of manual dispatching caused the slow development of shared electric vehicles in China.The emergence of autonomous driving technology is expected to realize the autonomous scheduling of shared electric vehicles and achieve a huge transition from “people looking for cars” to “cars looking for people”.The shared autonomous electric vehicles(SAEVs)service will integrate the environmental and social benefits of electric vehicles and shared vehicles,and provide a more intelligent and flexible demand response service than traditional car-sharing.This paper intends to study related routing optimization problems in SAEVs services from the operational perspective,combined with mathematical modeling and solution algorithm design.Firstly,in view of the existing charging technology,the routing problem of SAEVs under different charging strategies was studied.Considering the facility decision-making problems faced by SAEVs in the early stage,this study integrated and optimized the facility location,the fleet size,and the vehicle route.The facilities serve both as depots and as charging facilities.A nonlinear SAEVs routing optimization models under different charging strategies were formulated considering partial charging,full charging,and battery swapping.A real-numberencoded genetic algorithm combined with simulated binary crossover and polynomial mutation operator was developed to solve the problem.The non-linear programming solver from GAMS software and the genetic algorithm were used to solve the instances of different sizes,which verified that the proposed genetic algorithm can solve the model efficiently.Comparing the optimization results under different charging strategies shows the economic superiority of battery swapping in optimizing the vehicle routes.Parameter sensitivity analysis shows that using vehicles with larger battery capacity and setting smaller minimum battery level ratios can encourage operators to operate smaller fleets with lower costs.Secondly,in view of the high efficiency of speed control of autonomous vehicles and the influence of variable speed on energy consumption,the routing problem of SAEVs considering speed optimization was studied.Combined with the application scenarios of SAEVs using public charging facilities in the future,this study integrated and optimized the vehicle service routes and the driving speed.Battery swapping was used as the replenishment method for SAEVs due to its economy.A mixed-integer linear model integrated with speed variable decision for SAEVs route optimization was formulated.And an adaptive large neighborhood search(ALNS)algorithm was designed to solve the problem.A speed optimization procedure was used to optimize the driving speed of the vehicle while arranging the charging plan.The CPLEX solver and the ALNS algorithm were used to solve the instances.The ALNS shows good performance compared to the best-known solutions.The comparison of whether to make speed decisions during route optimization shows the economic benefits of speed optimization,which is mainly reflected in the saving of travel time and energy consumption.Furthermore,the positive correlation between energy consumption and travel distance,the contradictory relationship between energy consumption and travel time,and the dependence of different goals on speed were explored.Parameter sensitivity analysis shows that larger battery capacity and smaller minimum battery level ratio settings can help reduce the vehicle’s travel distance,while smaller battery capacity and the feasible lower speed can help save energy consumption.The travel time of the vehicle only depends on its weight in the objective.Finally,in view of the traffic congestion,the routing problem of SAEVs considering departure time and speed optimization was studied.It integrated and optimized the vehicle route,the departure time,and the driving speed.A mixed-integer linear model for SAEVs route optimization that integrates the decision of departure time and speed variables was formulated.The ALNS algorithm embedded with the departure time and speed optimization procedure was designed to solve the problem.The CPLEX solver and the ALNS algorithm were used to solve different-sized instances.The results show that the designed algorithm can effectively solve the model.The comparative experiment of the instance shows that the optimization of departure time can effectively avoid the congestion period of vehicles,and speed optimization also helps to save costs.From the parameter sensitivity analysis,it is found that the longer congestion time mainly increases costs by affecting travel time,and the lower average vehicle speed caused by congestion can contribute to the reduction of energy consumption.The battery capacity and the minimum battery level ratio have a greater impact on the travel time of passengers.From the perspective of travel time,if operators use vehicles with larger battery capacity and set a smaller minimum power ratio,it can be conducive to improving the quality of travel services.