Distribution Routing Optimization Based On Mobile Phone Data Mining

Distribution is one of the key scientific issues in the logistics,and is a key problem in logistics management and service.With the intensification of the competition and the slowdown in economic growth,the business volume and the profit of distribution are both on a downward trend.These compel the companies to introduce new technologies for the vehicle scheduling and the distribution management to improve the operational efficiency,the customer satisfaction and the core competitiveness.Dynamic Vehicle Routing Problem(DVRP)is a research field that has been attracting attention for more than ten years.In actual distribution,information such as demand,travel time and service time may be uncertain.With the development of computer technology and information technology,try to establish the mathematical model of vehicle routing problem based on the priority prediction of dynamic information through data mining.Aiming at the common factors in DVRP,the dynamic traffic network and dynamic demand VRP are studied respectively in this paper.First of al,through the study of travel time acquisition technique and mobile phone spatio-temporal data structure,this paper designs a method to obtain the travel speed by using the spatiotemporal data mining of the mobile phone.The BP neural network is used to predict the travel speed of the road network to realize the speed-dependent function fitting.Using GY city four days of mobile phone spatio-temporal data mining data obtained by training the model to predict the city's road network travel speed,the realization of road network speeddependent function of the state of the road network fitting.The research lays the foundation for the research of the path optimization problem of the time-varying road network in the back and provides a new idea for the acquisition of the travel speed of vehicles with time-varying network.Secondly,the shortest path problem of time-varying road network is the foundation and key to solving VRP problem.Therefore,the shortest path model of time-varying road network is constructed.Considering the complexity of road network and calculation time,A * algorithm is used to solve the shortest travel time path between two points.In GY city a total of 369 nodes and 549 sections of the time-varying network,select 5 OD pairs and calculate the shortest travel time of the variable-pitch network under different departure times.Since the actual VRP problem requires a large number of calculation of the shortest path between two points,an efficient time-varying shortest path algorithm will provide a good foundation for speeding up the calculation of the problem.Thirdly,based on the uncertainty of road network performance,the travel speed of road segments changes with time and the travel time of each segment changes accordingly,establish vehicle routing model with time window.According to the basic situation of Jingdong distribution service,particle swarm optimization is used to optimize the distribution route of 33 stations using 1000 times iteration of 30 particles with 8 time functions.Then the path optimization results under different total demand are analyzed and compared from the objective function,the number of vehicles used,the total delivery time and the unit cost.Finally,considering the demand uncertainty of distribution site and the uncertainty of the performance of road network,an optimization model of vehicle routing under time-varying road network under dynamic demand is established.Transform the uncertainty of the future into a static one by travel speed dependent function and estimated demand through time segment.In the meanwhile,the update strategy of particle swarm optimization is introduced into the frog leaping algorithm.The improved frog leaping algorithm is used to realize the problem under the condition of 33 sites and dynamic change in 8 time periods.At the same time,the optimization effect under different unit penalty costs of unfinished demand is studied.The results show that when the unit penalty cost is higher,the model tends to meet more demand to reduce the penalty cost.The demand increase also leads to the increase of delivery vehicles.The model and solution algorithm can solve practical problems and guide the operation of enterprises and improve work efficiency.