Research On Proactive Vehicle Routing Problem Based On Demand Forecast

In recent years,the level of urban logistics has risen steadily,which has a huge positive impact on the improvement of residents' quality of life,and the gradual improvement of consumption level also has higher requirements for logistics service quality and service efficiency.It is urgent for logistics distribution enterprises to upgrade.Not only the logistics industry focuses on it,but also the academia pays more attention to the logistics distribution mode.With the development of high-tech technology and the skilled application of big data's technology,the proactive vehicle routing distribution mode,which takes the data-driven mode as the core,brings new opportunities and development for urban distribution.Proactive vehicle scheduling is an integrated optimization solution driven by customer historical data and combined with relevant theories in other fields,which has strong robustness and strong antiinterference.It can evaluate and predict the uncertainty that may exist in each link before the project implementation,and at the same time deal with the unexpected dynamic events occurred in the process of project implementation,so that it can quickly return to normal operation.Compared with the traditional vehicle scheduling based on passive response,the proactive vehicle scheduling method can not only achieve efficient and high-quality urban distribution,but also significantly reduce the distribution cost.Firstly,this paper introduces the significance and background of the topic,and then expounds the current research situation from three aspects: dynamic problem research,proactive prediction guidance research and routing algorithm research.Secondly,it expounds the theoretical concept and lays the groundwork for the related research.Through the analysis of the current traditional distribution mode,this paper leads to the thinking of the problems existing in the distribution mode,and carries on the related research by combining the demand forecast and the dynamic vehicle distribution problem.The distribution process includes known static customers and dynamic customers with uncertain demand.Through the analysis and verification of customer historical demand data in order to predict dynamic customer demand in advance and set up distribution routes reasonably to achieve the purpose of minimizing the overall distribution cost,a simulated annealing-genetic algorithm is proposed to solve this problem,and a random Solomon standard example is used to verify the effectiveness of the algorithm in MATLAB.Thirdly,considering the disadvantages of single-stage distribution comprehensively,on the premise that the distribution historical data are available,the designed demand forecasting method is used to evaluate the uncertain demand in the distribution process.According to the evaluation results,service regions are divided and customers are clustered,to build a two-level logistics distribution network including hub logistics center,distribution logistics center and customer three types of nodes.Then the CW-Tabu search algorithm is designed to test the scale of medium-sized data,and the algorithm can be solved quickly and effectively.Finally,taking a large logistics distribution enterprise in Chongqing as an example,it is verified that the proactive demand quota strategy can schedule parts for its service more quickly from the belonging distribution logistics center.The results show that the adaptability of the single vehicle scheduling scheme to the traffic restriction policy is weaker than that of the 2E-VRPRF scheme,and the strategy proposed in this paper can better adapt to the traffic strategy and fully meet the needs of theory and practice.