Effects of mutation and crossover operators in the optimization of traffic signal parameters

Abstract

In this work, we analyze crossover and mutation operators for traffic signals optimization aiming to understand the problem from a system level perspective. We use MATSim to simulate the transport system of the business district of Quito (Ecuador) with 20000 agents moving in a one-day congested scenario. We relax the usual assumption of common cycle length for all signals and minimize travel time focusing on the optimization of 11 consecutive signals located in a main road. We study individual and combined effects of crossover and mutation for cycle length, offset, and green times. The results of this study provide valuable insights to know better the problem, validate the mobility scenario, and understand the effects of the operators.