Research And Design Of Carpooling Based On Intelligent Algorithm

With the rapidly development of the urban construction and the automobile industry, the contradiction between the urban traffic supply and the urban traffic demand is increasingly prominent. In order to relieve the pressure in the city, the mode that some people share one car is paid more attention by people. So the path optimization algorithms and the car combined algorithms become a hotspot. Currently, most websites and mobile phone software about carpooling only offer car information and provide the supply and demand platform, but do not supply the matching function between passengers and vehicles. These websites and software don’t achieve the target witch is optimizing the path of carpooling.Based on the study of carpooling problem’s related theory and information technology, the paper combines the intelligent processing algorithms with carpooling problem to discuss the algorithm of solving vehicle ride matching problem and carpoolingroute optimization problem,explore particle swarm optimization and genetic algorithm in the application of carpooling by the algorithm principle, steps and characteristics etc.For multi-vehicles carpooling problem with soft time windows, setting up mathematical model of the shortest route and least cost as the objective function, and take appropriate measures to deal with the timeout. To solve multi-vehicles carpooling problem,first matching vehicles and passengers, planning the carpoolingroute with the results of matching. In solving the carpooling matching to make the matching rate as the target function, using particle swarm optimization to optimizethe search radius of the stations which vehicle has passed, making the vehicle and the passenger has the highest matching rate.Carpool matching is completed, using the improved genetic algorithm to optimize the carpoolingroute.For this research, from the initial population, selection strategy, crossover and mutation operator of these four aspects to improve genetic algorithms, accelerate the convergence of the algorithm, and more likely to get the optimal solution. In order to validate the algorithm, this paper designs a road network, a number of vehicles and passengers in the road network, each driver has an initial route and time windows,passengers have the appropriate time windows and terminal. The experimental results show that the particle swarm algorithm can effectively solve the problem of carpool matching, the mathematical model and the improved genetic algorithm can effectively solve the optimizationproblem of carpoolingroute.Finally this paper establishes a carpooling system based on android platform, the client of systemis mainly responsible forinteraction with user and interface display, the server of system is mainly responsible for using optimization algorithms to achieve carpool matching and carpoolingroute optimization. User login system release initial route and time window, the system uses particle swarm algorithm and the improved genetic algorithm matching drivers with passengers and carpoolingroute optimization, the matching results and the optimal routeshow on the client, verify the rationality of theoretical model and algorithm design.