Minimizing Emission for Timely Truck Transportation with Adaptive Fuel Injection

We consider the problem of optimizing truck operations, where heavy-duty trucks haul freights that requires timely delivery. The trucks are equipped with engines that intelligently switches among multiple fuel injection strategies for efficient emission profiles. Our objective is to minimize the emission, by optimizing both routing and speed planning while accommodating that modern combustion engines come with multiple control strategies switching at runtime based on the engine speed. Besides the combinatorial nature of simultaneous routing and speed planning that already makes the problem NP-complete, the problem also face the unique challenge that the emission is non-continuous, non-convex with respect to the vehicle speed. We identify a special structure in the problem to provide an analytical solution for the optimal speed profile, and then develop an efficient heuristic for both path planning and speed planning to deal with problems on the scale of national highway systems. We perform simulation on the US highway system, and demonstrate that our approach reduces on average 18% emission compared to the baseline with a single control strategy. Compared to the fastest path approach that is adopted in common practice, our scheme reduces 72% emission on average.