An effective EA for short term evolution with small population for traffic signal optimization
Rolando Armas, Hernan E. Aguirre, Fabio Daolio, Kiyoshi Tanaka · 2016
In this work, we study the effects of mutation operators combined with a varying mutation schedule applied to traffic signal optimization. An evolutionary algorithm with specialized mutation operators coupled with a microscopic traffic simulator tackles the optimization of traffic signal settings in different mobility scenarios. Experimental results show that the proposed mutation operators allow for an effective search in large decision spaces, evolving small populations for a short number of generations. The parameters of the evolutionary algorithm are analysed and automatically-generated configurations are discussed, suggesting alternative ways to effectively apply the proposed variation operators for short term evolution.