Fuel-optimal truck path and speed profile in dynamic conditions: An exact algorithm

We consider optimizing a truck's choice of path and speed profile to minimise fuel consumption, exploiting real-time predictive information on dynamically varying traffic conditions. Time-varying conditions provide particular challenges, both from network-level interactions (e.g. slowing consume more locally may be beneficial avoid congested periods downstream) link-level phenomena interaction between acceleration gradient profiles). A multi-level, discrete-time decomposition the problem is presented in which: (i) [sub-problems] profiles are optimized within each link, given boundary entry/exit times speeds; (ii) [master problem] space-time extended network representation used encode dynamic interactions, which joint made. By instantiating with optimal link (i), we able devise tractable algorithm while over fine timescale. The solution approach pre-solve offline computationally-intensive step meaning that can efficiently produced online response information, whereby optimization solved by single shortest search network, for many exact algorithms exist. method additionally discretionary stops (pre-trip) departure time. Two representations investigated, depending whether constraints imposed ensure consistency across boundaries. Numerical experiments reported small illustrative example case-study network.