Efficient Mandatory Lane Changing of Connected and Autonomous Vehicles

In a mandatory lane-changing scenario, vehicles need to move to their target lanes before the end of a road segment. Mandatory lane changing usually happens near a highway ramp, and it is one major source of traffic congestion and delay. The advance of connected and autonomous vehicles supports to solve the problem in a centralized and precise way. We formulate a mandatory lane-changing problem and propose a Mixed-Integer Linear Programming (MILP) approach to find an optimal solution. We then propose an MILP-based incremental-window algorithm to solve the problem efficiently. Experimental results show that the MILP-based incremental-window algorithm can take much less computation time and achieve almost the same solution quality, compared with the original MILP approach. The advantage of efficiency yet sufficient effectiveness matches the need of high-level control (decision of passing order) well.