Abstract: To pay equal attention to efficiency, fairness, and spillback in the process of network traffic control remains an open problem. A multi-intersection coordinated control model with equal emphasis on efficiency, fairness and spillback is constructed by using a rule-driven method that combines fuzzy logic, genetic algorithm, and multi-agent system, where the traffic signal of each intersection is controlled by an agent that contains four modules, the information collection and distribution module is responsible for collecting and sending traffic condition at local and adjacent intersections, the phase selection module is responsible for selecting the next green phase, the green time calculation module is responsible for calculating the green time for the selected phase, and the decision optimization module is used to optimize the fuzzy rules and membership functions of both the traffic urgency fuzzy inference machine of the phase selection module and the green time fuzzy inference machine of the green time calculation module. Experimental results show that optimizing the aforementioned fuzzy inference machines can significantly improves the performance of the model, independent control with equal emphasis on efficiency and fairness is more efficient and fairer than fixed-time control and vehicle-actuated control, and coordinated control with equal emphasis on efficiency, fairness and spillback is not only more efficient and fairer than independent control, but also effectively prevents spillback, and makes the distribution of network traffic flow develop in a more balanced direction.

Key words: network traffic; coordinated control; efficiency; fairness; spillback