A multimodal traffic responsive strategy using particle swarm optimization

The paper describes a model predictive control strategy for multimodal urban traffic regulation for private vehicle (PV) traffic and public transport (PT). By acting on the traffic lights, the strategy aims to achieve two objectives. First to make public transport vehicles respect their timetable and, second, to improve the global traffic conditions. Based on a complex model, the strategy determines the optimal green splits and the offsets of the traffic lights. The optimization problem is multiobjective and resolved by the particle swarm optimization algorithm. Two criteria are proposed for private vehicle traffic and public transport. One of the originalities of this work is that the two criteria are expressed globally and are related since the PT progression model depends on the PV traffic. Two PSO algorithm versions are used: the original algorithm and a modified version expected to be more efficient. The efficiency of the strategy is evaluated on a large virtual urban network.