Steady State Traffic Optimization by MITROP

Abstract MITROP (Mixed-Integer TRaffic Optimization Program) is a new computer program for traffic signal control in urban street networks. It is designed to optimize simultaneously all the traffic control variables of the network including cycle time, splits of green time, and offsets. Both deterministic and stochastic models are used to represent the traffic flow process. The optimization problem is formulated in terms of mixed-integer linear programming and a globally-optimal solution is determined via the MPSX optimization system. The method was applied to several signal-controlled traffic networks, for various demand patterns, and offers certain advantages over existing methods. In particular, it is shown that network performance is sensitive to variations in cycle time, which has to be determined in conjunction with all the other traffic control variables. The paper describes the application of the program to a section of the UTCS/BPS test network in Washington, D.C.