The heterogeneous green vehicle routing and scheduling problem with time-varying traffic congestion

• A new mathematical model is proposed to reduce CO2 emissions for vehicle routing.
• A combination of exact and heuristic methods is used to solve large-sized problems.
• Considering traffic congestion in vehicle routes and schedules reduces emissions.
• Distance/time-based schedules do not always reduce emissions under traffic congestion.
• The proposed model is tested on simulated real-life cases.

The green vehicle routing and scheduling problem (GVRSP) aims to minimize green-house gas emissions in logistics systems through better planning of deliveries/pickups made by a fleet of vehicles. We define a new mixed integer liner programming (MIP) model which considers heterogeneous vehicles, time-varying traffic congestion, customer/vehicle time window constraints, the impact of vehicle loads on emissions, and vehicle capacity/range constraints in the GVRSP. The proposed model allows vehicles to stop on arcs, which is shown to reduce emissions up to additional 8% on simulated data. A hybrid algorithm of MIP and iterated neighborhood search is proposed to solve the problem.