Research Of Bus Dispatch Optimizing Based On Parallel Computing

At present, in large and medium-sized cities, the focus of the public traffic management is the public transportation scheduling problem. With the rapid development of urban and rural construction, urban scale has been expanded, the traditional operation mode no longer applies to the transit truth now, it requires the urban public transport which can be more quickly. In order to meet the demand of people, bus companies have to optimize the design of bus scheduling method. The core of bus scheduling is to work out a plan for a reasonable start using the scientific scheduling algorithm.Before making the scheduling plan, it need to collect real-time traffic information and vehicle information as scheduling base, GPS/GIS system, wireless communication technology and wireless radio frequency technique provide the technical support for information gathering. Each technology module is integrated into the bus vehicle terminal, by positioning system and passenger IC statistical device, the terminal get the vehicle information and passenger information respectively, and then, the real-time information is sent to scheduling command center through wireless communication network. After scheduling command center receive real-time information, it analyzes the information and sorts out reliable data used for the bus scheduling.In solving the path optimization and the job scheduling, artificial intelligence algorithm is of high efficiency, therefore, research on choosing the appropriate artificial intelligence algorithm, using which to work out bus scheduling optimization problems, has the realistic basis and theoretical feasibility.The Ant colony algorithm is selected as bus scheduling algorithm, enterprise cost and passenger time cost for bus are made as constraint conditions, the bus scheduling model is converted into a mathematical model, designed the objective function for scheduling algorithm. By studying the combinatorial optimization of departure interval in different periods, the minimum value of objective function is got finally.This paper proposes using the parallel computing method combined with ant colony algorithm for scheduling. Ant colony algorithm has parallelism, and the advantage of parallel computing is that using space for time and parallel processing of multiple processes, finally improve the efficiency of operation. Using parallel ant colony algorithm overcome the defect that single use of ant colony algorithm is of low operation efficiency and bad accuracy. By comparing a lot of experimental data, it is determined that the optimal parallelism of parallel ant colony algorithm program, which is suitable for the public transportation scheduling problem.Paper also analyzes the deficiency of using parallel Ant Colony Algorithm for bus scheduling, and the future work direction and focus are proposed, the improved parallel ant colony algorithm will have broader prospects in bus scheduling.