Research On Modeling And Optimization Of Large-scale Dynamic Ridesharing Systems

Dynamic ridesharing matches the vehicle and the passenger together in real time,and plans the optimal routing for the vehicle.Through certain pricing strategies,the cost of taking taxis can be reduced for passengers or lower than the original travel cost for drivers to stimulate passengers and drivers to participate in ridesharing.Dynamic ridesharing increases the number of passengers carried by vehicles and reduces the number of vehicles running in the transportation system.It is an effective means to ease traffic congestion and is increasingly favored by people.In this paper,considering the preference of acquaintances,the same sex and the maximum number of passengers allowed to carry,the matching model based on preferences is established.By adjusting the relative weight parameters in the objective function,the model achieves the goal of maximizing the rate of ridesharing,minimizing the cost and maximizing the preference.In order to solve the model,we design a heuristic algorithm,Greedy Random Adaptive Search Algorithm(GRASP).Finally,an example is used to verify the validity of the model,and the results show the correlation between the parameters and the corresponding impact variables.By adjusting the proportion of female passengers in the input data,the influence of different sex ratios on the matching rate is analyzed.The results show that the model can deal with the matching of the passenger and the vehicle with the acquaintance,the same sex and the preference of the maximum number of passengers.For ridesharing participants,unreasonable pricing leads to the inability of passengers and drivers to reach a deal,and affects travel.For the ridesharing system,unreasonable pricing reduces the success rate of matching and increases the system cost.we design a dynamic ridesharing pricing strategy based on detour.In this strategy,the ridesharing pricing was calculated in each section based on the participants’ original travel distance and detour distance after the system segmented the vehicle routing.A decomposition algorithm for large-scale network is designed to deal with the matching problem,and dynamic programming algorithm is used to solve the vehicle routing problem in this paper.The idea of the algorithm is to decompose the large-scale problems into independent small-scale problems based on the pre-processing results,and solve these small-scale problems by parallel computing.Finally,the numerical results show that this pricing strategy can significantly improve the matching rate and reduce the system cost.