Optimal signal timing signalized intersection by global optimization (OPT-I)

Abstract

Accurate measurements of signal control parameters at signalized intersections are very important for designing and operating traffic control systems. In this work, a queuing system has been modelled as a result from the characterization of the vehicles behaviour approaching and passing of a signalized intersection. This intersection is part of a network of urban traffic and two problems are formulated, distinguished by the type of control used: pretimed, semi-actuated or fully-actuated. Subsequently global optimization and complementarity can be used to determine the parameters of the control signal. This formulation includes the green times and cycle lengths that minimize the total waiting time of vehicles at the intersection. At intersections regulated by actuated control, this methodology allows us to estimate the effective green time in actuated streams as well as the length of each cycle. It should be noted that the duration of departure of vehicles when signal is green or yellow is constant. The vehicles arrivals are random, following a Poisson probability distribution. The models in question were formulated as linear programs with linear complementarity constraints (LPLCC). In the present study, the Sequential Linear Complementarity Algorithm (SLCP) was also analysed to calculate the global minimum for the LPLCC. Furthermore, several scenarios of traffic intersections are created to demonstrate the method's efficiency, particularly to verify the accuracy of the solutions of the problems.