An Agent-Based Simulation of Truck Fleet Operations to Supply a Biorefinery Year-Round

Background: The cost to operate a truck fleet to haul feedstock from satellite storage locations (SSLs) to a biorefinery is typically more than 30% of the logistics cost. The use of central control can minimize truck wait times and maximize truck productivity (Mg hauled per day). Methods: An agent-based model, developed in Python 3.12.7, simulated truck hauling operations and SSL loading operations at a 1 min time step for each day in a six-day workweek over a 48-week hauling season for a theoretical biorefinery centered in Gretna, VA, USA. Several truck and SSL operational parameters included stochastic components to allow for random variability (e.g., drive speed and loading rate). Results: The theoretical minimum fleet, assuming no unproductive time, was 6 trucks. Assuming realistic delays, a fleet of 13 trucks could supply the biorefinery with one unloading operation, but unproductive time was over 50% of the hauling day. Conclusions: By adding a second unloading operation at the biorefinery, average truck idle time was reduced, and a fleet of 8 trucks could achieve the same average truck productivity with total unproductive time reduced to about 20% of the total truck fleet operating time.