Urban Distribution Path Optimization Base On Real Time Traffic Information

Recently,logistics industry is development rapidly.However there are still big challenges for distribution companies,on the one hand,with a crazy urban environment,the customers wish to receive the goods under a certain time,higher requirement are put forward for the distribution limitation;On the other hand,urban traffic is complicated and changeable,the phenomena of the traffic congestion become more and more serious which affects the timeliness of urban distribution.therefore Precision delivery service becomes the main direction of logistics enterprise development.With the development of big data and information technology,widespread application of GPS\GIS,we can provide real-time and concrete road information for travelers.So use urban real-time road information to further planning a reasonable travel path.has become an urgent solved problem.This paper considers the influence of on-line road information on the vehicle routing,and also considers the customer time window,constructs a hybrid algorithm model to solve this problem.For real-time traffic information,we make advance prediction according to historical data and dynamic adjustment.The update strategy is designed for the time-varying network which update date at customer points and rode nodes.the result of time window function directly affect the concentration of updating the pheromone.This model combined ant colony algorithm and Dijkstra algorithm.Illustrated the design process of hybrid algorithm and operate process.For ant colony algorithm,change transfer probability in order to void falling into local optima.For Dijkstra algorithm,we improvement the week point of in the problem of previous adjacent vertice in shortest path by using the broken circle approach.Finally window7 as the development platform with AnyLogic7.3.7 as the development tool.through the simulation platform constructed the algorithm model by using Agent model ideology with Java.the result of the example based on actual road network demonstrate that the above model and algorithm are effective.