Truck Platoon Formation at Hubs: An Optimal Release Time Rule

TL;DR

This paper develops an optimal stopping rule for coordinating truck platoon formation at hubs, balancing benefits and waiting costs, and demonstrates near-optimal performance compared to idealized non-causal strategies.

Contribution

It introduces a monotone stopping rule model for hub-based platoon coordination and derives a simple one-step look-ahead optimal policy.

Findings

Optimal rule closely matches non-causal performance.

Outperforms periodic release strategies, especially at low arrival rates.

The policy is simple and computationally efficient.

Abstract

We consider a hub-based platoon coordination problem in which vehicles arrive at a hub according to an independent and identically distributed stochastic arrival process. The vehicles wait at the hub, and a platoon coordinator, at each time-step, decides whether to release the vehicles from the hub in the form of a platoon or wait for more vehicles to arrive. The platoon release time problem is modeled as a stopping rule problem wherein the objective is to maximize the average platooning benefit of the vehicles located at the hub and there is a cost of having vehicles waiting at the hub. We show that the stopping rule problem is monotone and the optimal platoon release time policy will therefore be in the form of a one time-step look-ahead rule. The performance of the optimal release rule is numerically compared with (i) a periodic release time rule and (ii) a non-causal release time rule where the coordinator knows all the future realizations of the arrival process. Our numerical results show that the optimal release time rule achieves a close performance to that of the non-causal rule and outperforms the periodic rule, especially when the arrival rate is low.