Abstract
This dissertation documents the development of optimization models in a mixed integer-linear form for the control of network traffic signalized intersections. The existing network traffic signal optimization formulations usually do not include traffic flow models, except for control schemes such as SCOOT that use simulation for heuristic optimization. Other conventional models normally use isolated intersection optimization with traffic arrival prediction using detector information, or optimization schemes based on green bandwidth. In this dissertation a complete formulation of the problem that includes explicit constraints to model the movement of traffic along the streets between the intersections in a time-expanded network is presented, as well as constraints to capture the permitted movements from modern signal controllers. The platoon dispersion model used is the well-known Robertson’s model, which forms linear constraints. Thus it is a rare example of a traffic simulation being analytically embedded in an optimization formulation. The formulation is an integer-linear program, and does not assume fixed cycle lengths or phase sequences. It assumes full information on external inputs, but can be incorporated in a sensor-based environment. The integer-linear program formulation may not be efficiently solved with standard simplex and branch and bound techniques. We discuss network programming formulations to handle the linear platoon dispersion equations and the integer constraints at the intersections. A special purpose network simplex algorithm for fast solution is addressed in the proposed solution approach. The optimization model takes the form of mixed integer linear programming. The control strategies generated by these optimization models were compared with those derived from conventional signal timing models, using the TRAF-NETSIM microscopic simulation model. It was found that the optimization models successfully produced optimal signal timing plans for the various signalized intersections including simulated and real-world networks. The proposed optimization models consistently outperformed the conventional signal control methods with respect to system delay objective. This conclusion was drawn from the TRAF-NETSIM simulation.