A real-Time ride-sharing matching framework using simulated annealing genetic algorithm

Abstract

Recently, ride-sharing services are becoming more popular in commercially application, which have attracted many researchers' attention. The ride-sharing provides significant economic, societal and environmental benefits in a sharing economy, such as reducing pollution, travel costs and traffic congestions. This kind of problem essentially is to maximize the matched ride-sharing pairs, and obviously is an optimization problem. There are lots of platforms for dynamic peer to peer ride-sharing, however, existing researches can be improved better. For example, how to reduce the algorithm computing time, and maximize the matching effectiveness to gain more economic benefits. In this work, we first introduce the problem of ride-sharing with time guarantee on road network, and then design a novel heuristic simulated annealing genetic algorithm. In addition, we carefully adjust the parameters with different constraints, and conduct extensive verification experiments with realistic datasets derived from Beijing car services, the results demonstrate the advancement of our methodologies.