Bit level versus gene level crossover in a traffic modeling environment

Abstract

During a couple of years we have dealt with the Traffic Lights Optimization Problem. We have designed a scalable and flexible architecture for this purpose. By this work we present a comparison between two different aproaches to crossover We have used a real world case. Through some tests we show that using a gray code bit level crossover instead of a gen level crossover may suppose a great saving of computing power.