Hybridization of Racing Methods with Evolutionary Operators for Simulation Optimization of Traffic Lights Programs.

In many real-world optimization problems, like the traffic light scheduling problem tackled here, the evaluation of candidate solutions requires the simulation of a process under various scenarios. Thus, good solutions should not only achieve good objective function values, but they must be robust (low variance) across all different scenarios. Previous work has revealed the effectiveness of IRACE for this task. However, the operators used by IRACE to generate new solutions were designed for configuring algorithmic parameters, that have various data types (categorical, numerical, etc.). Meanwhile, evolutionary algorithms have powerful operators for numerical optimization, which could help to sample new solutions from the best ones found in the search. Therefore, in this work, we propose a hybridization of the elitist iterated racing mechanism of IRACE with evolutionary operators from differential evolution and genetic algorithms. We consider a realistic case study derived from the traffic network of Malaga (Spain) with 275 traffic lights that should be scheduled optimally. After a meticulous study, we discovered that the hybrid algorithm comprising IRACE plus differential evolution offers statistically better results than conventional algorithms and also improves travel times and reduces pollution.