Traffic Control Through Traffic Lights Management: A Comparison Study

Traffic jams are a severe problem affecting cities around the world, and they are only getting worse as the population and number of vehicles continues to increase significantly while the area available for building roads does not. Annually, many citizens waste countless hours stuck in traffic, and the amount of time spent is expected to increase. One study calculated that in 75 different city regions in 1982 the average number of hours wasted per year per traveler during peak hours was 16. By 2000, this had increased to 62 hours [1]. However, in spite of this, many roads are rarely congested. As such, one approach to dealing with this is solving the Urban Traffic Control Signal problem. Simply put, the traffic lights of the city could be used to redirect traffic such that vehicles are moved away from heavily congested roads and onto less congested ones, thereby reducing the amount of time wasted stuck in traffic. The problem then is figuring out how to generate appropriate signal timing patterns, as most simple patterns are not sufficient to optimally redirect traffic and reduce congestion. Thus many researchers have actively employed different approaches to deal with the problem. Out of these approaches, there have been three main categories: Network Flow approaches, Reinforcement Learning approaches, and Game Theory approaches.