Real-time metaheuristic-based urban crossroad management with multi-variant planning

This paper presents a multi-variant planning method for the problem of multi-lane crossroad management. The method leverages a metaheuristic system which is aimed at real-time usage. Being the basis of the experiments shown, its implementation is scalable and can efficiently use a basic multi-core hardware infrastructure. The whole system can provide a sub-optimal yet useful crossroad management plan and is perceived to be superior to the classic and competitive methods previously tested.