A better understanding on traffic light scheduling: New cellular GAs and new in-depth analysis of solutions

Vehicle traffic congestion is an increasing concern in metropolitan areas, with negative implications for health, environment, and economy. Researchers, city managers, and entrepreneurs have shown great interest in Smart Mobility, and several approaches have been proposed to reduce these non-desired effects. In this work, we focus on using the existing infrastructure (traffic lights) to tackle these negative issues, instead of investing in an expensive new one. The adequate planning of traffic lights (the configuration of the red-yellow-green cycles) improves vehicle flow (reducing jams, emissions, economic losses, etc.) and, at the same time, this improvement is obtained without any additional cost and without requiring the use of specialized applications by the drivers. We propose two versions of a Cellular Genetic Algorithm (cGA): synchronous and asynchronous. This method has previously shown very accurate results in real-world problems. Our approaches are evaluated with two closer-to-reality scenarios from urban areas located in the cities of Malaga (Spain) and Paris (France) using the popular micro-simulator Simulator of Urban Mobility (SUMO). A complex simulation of the city is mixed with an advanced (though light) algorithm to address a major problem in all cities. We compare our algorithm with respect to the state-of-the-art techniques for this problem, showing high accuracy of our techniques. Additionally, we present an in-depth analysis of the solutions obtained via a genotypic and phenotypic data science study, so that the whole domain gets a better understanding of what the algorithms are computing and experts can learn better strategies. (C) 2020 Elsevier B.V. All rights reserved.