Hyper-Heuristic Realtime Flows Guided Traffic Flow Assignment Algorithm

Since traditional traffic flow assignment methods are difficult to apply to real scenarios with large-scale road networks and uncertain traffic demand, this paper proposed a real-time traffic flow assignment algorithm based on genetic programming to break through the bottleneck of traffic flow assignment in dynamic open environments. The proposed method does not directly plan the path for each individual in advance but automatically evolves the routing rule for selecting the path by integrating the vehicle's own status and road network information. Under the guidance of this rule, individuals realize dynamic and real-time traffic flow assignments. The experimental results on the simulation platform for different urban road networks and traffic flows further demonstrate the effectiveness of this method compared with the comparative algorithm, and it has certain scalability and generalization capabilities.