Research On Takeaway Order Allocation And Path Optimization

In recent years,with the change of residents’ dining habits and the transformation and upgrading of the catering industry,the take-out market has developed rapidly,the number of users and the transaction scale have increased sharply,but the customers’ demand for shortening the delivery time is still increasing,and the distribution pressure of the platform is increasing day by day.Improving the distribution efficiency has become the key for the platform to maintain long-term competitive advantages.Take-out order allocation and route optimization,as two key links of take-out delivery,are the focus of research needed by the platform to reduce delivery costs and improve delivery efficiency.In view of these two links,the specific work of this paper is as follows:(1)The distribution and route optimization of take-out orders are summarized as take-out delivery problems for comprehensive research.Aiming at the characteristics of the delivery time of take-out orders,the research idea of first clustering,then assignment,and finally route optimization is proposed.To take delivery and take-out delivery problem with time Windows of dynamic vehicle routing problem(DVRPSPDTW)subset into order clustering problem and order assignment problem with pick up and delivery time Windows and multiple TSP problem,greatly reduces the complexity of the take-away distribution problem and solving the difficulty.(2)An improved K-Medoids clustering algorithm suitable for takeout order allocation is proposed.In the process of forming order subset by cluster analysis,the traditional K-Medoids algorithm is improved.Firstly,the polar coordinate scanning method is used to determine the initial clustering center to improve the convergence rate of clustering.Secondly,the concept of order distance is proposed and used as a clustering measure,which improves the application degree of the algorithm to the problem of takeout delivery.Finally,the balanced constraint conditions are added into the clustering algorithm to avoid the overload of riders and reduce the phenomenon of order timeout,which reflects the principle of balanced allocation.(3)The open test set of the outbound platform is taken as a case study of the instant order mode,and the effectiveness of the model and algorithm is analyzed.Comparative analysis and sensitivity analysis of static cases are carried out,and relevant parameters of the model and algorithm are adjusted.The results show that reasonable setting of merchant distance weight and space conversion coefficient is beneficial to reduce distribution cost and improve distribution efficiency.Then,the feasibility and robustness of the adjusted model and algorithm are verified by a dynamic example,and it is pointed out that the value of time slice should refer to the distribution of meal delivery time,and appropriate slice length will improve the order distribution efficiency and reduce the distribution cost.(4)Proposed solutions to the mixed order mode of adding reservation orders and analyzed the impact of reservation orders on takeout delivery.The pre-insertion method is proposed for the mixed order mode,which can improve the importance of the order in the path optimization process and ensure the accuracy of the order delivery.Subsequent experiments show that the pre-insertion method is effective,and it is pointed out that properly increasing the proportion of orders can improve the efficiency of distribution and reduce the distribution cost.This paper makes a comprehensive study on the distribution and route optimization of delivery orders,and its research results can provide reference value for the comprehensive study of delivery problems and provide decision support for the delivery platform.(4)Proposed solutions to the mixed order mode of adding reservation orders and analyzed the impact of reservation orders on takeout delivery.The pre-insertion method is proposed for the mixed order mode,which can improve the importance of the order in the path optimization process and ensure the accuracy of the order delivery.Subsequent experiments show that the effectiveness of the pre-insertion method is verified,and it is pointed out that properly increasing the proportion of orders can improve the efficiency of distribution and reduce the distribution cost.In this paper,an in-depth analysis is made on the distribution and route optimization of outbound orders.The research ideas can provide reference value for the subsequent research on this kind of problems and provide decision support for the platform.