With the rapid development of urban traffic and the increasing transportation demand, the traffic congestion
becomes more and more serious so that the time delay and economic loss are more and more big. The optimal signal timing
strategy is investigated for an urban intersection. According to the concept of queue length balanced controlling, the green
time for each phase is adjusted dynamically to achieve the equilibrium of queue length and make full use of green time.
Approximate dynamic programming is adopted so that the algorithm possesses the characteristics of self-learning and self-
adaptation, independent of the traffic flow model. The simulation results show the effectiveness of the proposed algorithm.