| In recent years,as customers have higher and higher requirements on the timeliness of delivery,as well as changes the influence of COVID-19 pandemic,the Instant Delivery,an emerging logistics form,have become more and more diverse.Instant Delivery does not only help consumption becomes fast and convenient,which sorely changes people’s consumption habits,but also provide a large number of jobs for the society;the most typical ones are the Instant Delivery Riders.The main job of a rider is to pick up goods from a merchant within the time required by the customer,and deliver the goods to the location designated by the customer in time.This order fulfillment process can generally be abstracted as a Pick-up and Delivery Problem with Time Window(PDPTW).How to efficiently and effectively solve this problem has been one of the hot research topics.This paper aims at solving and extending the research of PDPTW problem combining the particularity of instant delivery crowdsourcing mode.This paper first starts from discussing the characters of instant delivery market,especially the market size,history and development.The paper analyzes the advantages and disadvantages of the common rider employment models and figures out the advantages of the crowdsourcing employment model.The paper reviews related works on instant delivery and on crowdsourcing mode;it also reviews works on PDPTW,which is highly related to instant delivery industry.This article proposes that if the model for every step in the Instant Delivery process is optimized from the rider’s point of view,especially the path planning for regulating rider’s routes for delivery,a good bargain between the rider’s income and the platform’s order acceptance rate can be achieved.This paper then demonstrates that the proposed problem,which is the Instant Delivery problem under the crowdsourcing mode,can be transferred into a special kind of time-consuming picking and delivery problem,and establishes a mixed integer programming model for this problem,and proposes using the GRASP×ILS algorithm for solving such problem.To prove the efficiency and effectiveness of the proposed method,comparison between the proposed method and ALNS method,which has been proved to be more effective by some previous researches,is conducted on various kinds of publicly available datasets for pickup-and-delivery problem.To further demonstrate that the proposed mathematical model of pickup-and-delivery problem and the GRASP×ILS algorithm have a positive effect on the actual business indicators of instant delivery,a simulation experiment has been conducted from the perspective of platform order rate and rider’s actual income;the experimental results prove that the win-win between the platform and the rider can be achieved using such model and algorithm.This paper expands the previous research on the traditional path-planning problem in PDPTW on a theoretical point of view.Based on the research proposed in this paper about the real-time order grabbing scenario of crowdsourcing riders for Instant Delivery,a new pathplanning problem model is proposed,and an efficient solution is designed to solve the problem.From the perspective of practical application,the research in this paper possesses practical significance for the daily operations of Instant Delivery platform,and also help the algorithm design of such platforms.It can effectively improve the quality of service for costumers and also improve the delivery efficiency.The method can help to form a positive cycle that improve the profit and efficiency of the enterprise and at the same time increase the actual income of the rider can happen at the same time.This will significantly improve the efficiency of the overall social operation... |
