Incorporating driving cycle based fuel consumption estimation in green vehicle routing problems

Traditionally, vehicle routing problems (VRP) were solved with the objective of minimizing total distance traveled. The rationale behind minimizing distance was perhaps that fuel consumption depends on distance traveled. Fuel consumption depends on several other factors besides distance. Recently, several authors have focused on directly minimizing total fuel consumed considering load carried and/or speed. However, it is surprising that none have considered the significant effect of acceleration on fuel consumption. Since fuel cost represents a significant fraction of operating cost, incorrect fuel consumption estimation may result in suboptimal routes and schedules. Here, we estimate fuel consumption while considering the effect of load, speed, and acceleration. We achieve this by using driving cycles (speed-time profile of a vehicle) that can be easily obtained from Global Positioning System (GPS) data. Modified versions of several standard VRP instances are used to test the effect of estimated fuel consumption using driving cycles. Test results show that using driving cycles results in an average fuel savings of 8-12% compared to using average speed.