Non-even Adaptive Compatible Optimization Control for Urban Oversaturated Traffic Network

To solve the traffic signal control problem in urban oversaturated traffic network, the total delay is divided into the delay in the main links and the delay in the minor links, and the control problem is formulated as a conflicted multi-objective control problem. In this paper, a new non-even adaptive multi-objective compatible optimization control algorithm is proposed, multi-objective evolution algorithm NSGA-II is modified, and a non-even Pareto front spreading mechanism is proposed to obtain some special areas on the Pareto front and taken as the optimization tool for the multi-objective compatible control algorithm (MOCCA). The ideas of adaptive population mechanism and iterative control algorithm are utilized in the modified NSGA-II algorithm to increase calculation speed of the real time control algorithm. A stability preference selection strategy is proposed to obtain stable controller. The proposed multi-objective compatible control algorithm is used to solve the oversaturated traffic network control problem and tested under the simulation environment consisting of 7 junctions. The simulation result shows that the compatible control algorithm proposed in this paper can handle the oversaturated traffic network control problem effectively compared with the fixed-time control method.