Developing combined genetic algorithm - Hill-climbing optimization method for area traffic control

Abstract

This study develops a genetic algorithm with TRANSYT hill-climbing optimization routine, referred to as GATHIC, and proposes a method for decreasing the search space, referred to as ADESS, to find optimal or near-optimal signal timings for area traffic control (ATC). The ADESS with GATHIC model is an algorithm, which solves the ATC problem to optimize signal timings for all signal controlled junctions by taking into account coordination effects. The flowchart of the proposed model with ADESS algorithm is correspondingly given. The GATHIC is applied to a well-known road network in literature for fixed sets of demand. Results showed that the GATHIC is better in signal timing optimization in terms of optimal values of timings and performance index when it is compared with TRANSYT, but it is computationally demanding due to the inclusion of the hill-climbing method into the model. This deficiency may be removed by introducing the ADESS algorithm. The GATHIC model is also tested for 10% increased and decreased values of demand from a base demand. © 2006 ASCE.