The Study Of Urban Taxi Intelligent Dispatching Method Based On The GIS And Multiagent Theory

With the unceasing development of our urban economy, more and more traffic problems begin to show. Taxi, as an important urban transportation, due to its lack of scientific development and planning, is confronted with a high rate of empty crusing, a low operation profit and other issues. The key to solve these problems lies in scientific management, dispatching and a rational allocation of taxi transport capacity resources. A new taxi intelligent dispatching method is presented in this paper, which uses multiagent and GIS technology to make the method more efficient than the centralized dispatching.Based on the analysis of centralized taxi dispatch method, considering the problems of tasks allocation besides the calculation and graphical display of taxi optimal route, a new method of taxi intelligent dispatchling is proposed ,combining wuth the multiagent and GIS technology. Firstly, this method establishes a taxi dispatching zone with an asymmetric divide way of agent alliance. The zone is correspondent with the customers request scope in a one to one pattern. Then one packet of request messages are sent to the same number of taxi agents of the corresponding region from the dispatch center. An alliance of taxi agents is formed to allocate the requests. Based on the BDI model, An improved distributed cooperatively negotiation algrithm is discussed for solving the requests allocation. Secondly, the algrithm of fixing optimal path is researched within different time periods for obtainning the quality of the taxi's service. During the time of stable traffic, based on the history data, taxi agents improve the Dijkstra algrithm for calculating the optimal path using intelligent ideas without considering the time variation of traffic flow status. At the peak period of traffic flow , considering the time variation of traffic flow status, a improved genetic algrithm is used for calculating the optimal path established with multiagent techology. Thirdly, with establishing a GIS database of topology structure, the MapGIS is developed in the VB6.0 environment for realizing the optimal path algrithms at different time periods. And the optimal path is displayed on the monitor.Finally, in MITSIMLab environment, compare the centralized taxi dispatching method with the one based on GIS and multiagent according to an instance. Then it has been proved that the taxi intelligent dispatching method has better dispatching effect which is presented in this paper.