In this paper a method for the determination of optimum signal plan sequences for a network of signalized intersections is described. The method developed for the improvement of a fixed-time traffic control strategy is based on the Dynamic Programming technique. The results of its implementation in the Moscow traffic control system (START) are also presented. Nowadays, the majority of researchers are oriented towards real-time traffic control systems. However, we found it worthwhile to improve fixed-time traffic control strategy performance bearing in mind the following facts: (1) most area traffic control systems use a fixed-time strategy, at least during the first phase of system implementation; (2) this strategy remains in effect as the standby control strategy in case of the failure of traffic detectors (which are necessary for any of the more complex traffic control strategies). This paper presents the method we developed to improve the fixed-time control strategy, based on the Dynamic Programming technique.