Multi-Agent Intersection Management for Connected Vehicles Using an Optimal Scheduling Approach

Today´s transportation systems are facing numerous issues resulting from the increased travel demands and limited capacities of roadway infrastructure. As a potential intelligent transportation system (ITS) solution, multi-agent intersection management systems have recently received increased attention with the rapid advance in wireless communications and comprehensive vehicular technologies. Most of the proposed multi-agent system approaches take a FIFO (first-in first-out) approach to time-space occupancy scheduling. However, by also optimizing the departure sequence, greater global benefits are possible. In this paper, we propose a modified multi-agent system with optimal scheduling of Vehicle Agent´s (VAs´) departure times. Compared with the FIFO-based system developed in the authors´ previous work, the modified system can provide more system-wide benefits in terms of mobility, reliability and sustainability. Simulation studies have shown improvements in travel times, but more importantly an approximately 58% reduction in travel time variability and 49%-60% reductions in (partial) stops. These leads to potential benefits in fuel consumption and pollutant emissions, primarily by carefully designing VAs´ trajectories through the intersection.