Design and implementation of an urban traffic control system for public transport using multi-agent systems

In this research, we aim to design and implement an agent-oriented system to prioritize public transport, particularly buses in urban intersections. In our proposed system, we assume that the bus works on the basis of a timetable on every station. In addition, at the station that is just before the intersection, based on the amount of delay and the number of passengers (estimated by using innovative weight sensor), a priority value is assigned to each bus. Then, the bus requests the intersection agent its required green light time to cross the intersection. The intersection agent also uses the estimated number of vehicles, as well as bus priority value and then produce the best and optimized timing using genetic algorithms. In this research, we proposed a new multi objective fitness for our genetic algorithms where the goal is to reduce traffic as well as determination the minimum time for traffic lights. The advantages of our proposed method are as follows: using genetic algorithm, the calculation of the best timing for traffic lights and our new proposed fitness reduces the number of calling genetic algorithm for timing of the lights by about 30 percent compared to normal genetic algorithm. In addition, taking into account the number of vehicles at intersections also improves the overall traffic situation. Also, considering the number of the bus passengers as the bus prioritization factor improves passenger satisfaction and subsequently will be effective on reducing urban traffic. We compared our result with three traffic control systems for intersections including fixed time system, no priority system and prioritize systems that used only multi-agent systems. The results show that: our proposed method reduces bus delays about 80%, compared to the fixed time and without priority methods. In addition, using our proposed method is more effective on traffic than other priority method that used only multi-agent systems method.