Research And Realization Of Multi-Vehicles Routing For Traffic State Improvement Based On Evolutionary Game

With the increasing number of vehicles in the road network,the problem of traffic congestion has become more and more serious.The speed of traffic resources expanding is far behind the growth rate of vehicles.The rapid development of V2X communication and edge computing technology make the intelligent transportation systems viable.Based on these technologies,we can coordinate multiple vehicles with the communication technology and provide dynamic routing schedule with edge computing techonology.In this way,we can alleviate traffic congestion,and improve transportation experience.In past researches,most routing algorithms were designed for a single vehicle.The optimal driving route is calculated by evaluating the current road status.But these methods overlooked the impact of vehicles.One same route may be recommended to multiple vehicles,which eventually contributes to a deviation between the actual traffic state and the predicted traffic state.And the traditional centralized multi-vehicle routing algorithms mostly convert the multi-vehicle route planning problem into an optimization problem.By finding the global optimal solution,route palnning for multiple vehicles are achieved.However,in order to achieve the global optimization,some vehicles’expirence may be sacrificed.To fix the problem mentioned above,a distributed multi-vehicle cooperation routing algorithm based on evolutionary game theory is proposed.This algorithm recommends different route for different vehicles at the same time.And each vehicle participating in the cooperation obtains the best payoff with the strategy recommended by the algorithm.This method can not only balance the traffic flow,but also improve the driver’s compliance level.The proposed algorithm is divided into two parts.The first part foucus on multi-vehicle routing at a single intersection.By modeling vehicles that are going to pass through this intersection as a population game,the problem of multi-vehicle routing is transformed into a problem of finding the Nash equilibrium state of the game model.Then with the revision protocol,the static game model is transformed into evolutionary dynamics.And then the Nash equilibrium is calculated by finding rest points of these dynamics.The second part focouses on the coordination of multiple intersections.A two-layer traffic cooperative control system model is proposed in this paper.This model solves the problem of non-optimal routing caused by overlapping control area and limited cognition of a single game RSU.Finally,this paper evaluates the performance of the game-based multi-vehicle routing algorithm through simulation,and verifies the effectiveness of the proposed algorithm.