Research On Optimization Of Shared Bikes Deployment Path Under Different Time Periods

The appearance of shared bike effectively solves the last mile problem of users’ traveling.However,the distribution of bike resources presents irregular dynamic changes in time and space with users’ traveling,resulting in the phenomenon of “bikes stacking” and “no bike to borrow”.In the work of redistributing shared bikes resources and recycling damaged bikes,the operating enterprises have problems such as high deployment cost and unreasonable path planning.Aiming at the above problems,considering the difference of demand of shared bikes’ use in the flat-peak period and the high-peak period,this paper proposed a divided-period deployment model,studied on the optimization model of the deployment path under different time periods,and solved the model based on the improved Whale Optimization Algorithm(abbr.WOA).The specific research works are as follows:(1)Aiming at the problem of uneven distribution of time and space of shared bicycles,a divided-period bike deployment mode was proposed.And the deployment path optimization models under static and dynamic time periods were separately constructed.In the flat-peak period,with the goal of minimizing the cost of the enterprises,the vehicle routing optimization model under the integrated mode of deployment and recycling was constructed.In the high-peak period,by introducing the theory of time axis and key points,the dynamic path optimization problem was transformed into the single-depot path optimization model and the semi-open multiple depot path optimization model after updating the information.(2)The improved WOA was proposed to solve the model.Considering that the WOA is easily trapped into local optimization,this paper added nonlinear adjustment control parameters a and cross-variation operations based on the basic WOA to coordinate the exploration and development ability of the WOA.For the solution of static models,this paper used integer coding,and an initial population method based on problem characteristics was designed.Combining the crossover and mutation operators of genetic algorithm to enhance the diversity of the population,and promoted the algorithm to jump out of the local optimum.To solve the dynamic model,based on the static problem solving method,an initial population method based on key point was designed.And the key points were used as the nodes connecting the sub-provisioning cycle routes to ensure the integrity of the entire deployment route.(3)The divided-period path optimization model was verified through test cases.In the flat-peak period deployment case,the cost of deployment and recycling integrated mode was compared with the results under the mode of “separate deployment + separate recycling”,which proved the feasibility of adopting the integrated mode of deployment and recycling in the flat-peak period.In the high-peak case,the model before the update of the deployment information and the model after the update of the deployment information were solved separately,and the lower limit setting of satisfaction was analyzed.Finally,compared with the dynamic deployment results with recycling,the rationality of the proposed dynamic without recovery mode was proved.In the static and dynamic models,the improved WOA,the basic WOA and genetic algorithm were compared to verify the effectiveness of the improved WOA.