Adaptive Route Optimization for Dynamic Fleets: A Hybrid ILS-SA Approach to the Vehicle Routing Problem with Occasional Drivers

Abstract

This study proposes a hybrid Iterated Local Search and Simulated Annealing (ILS-SA) algorithm for optimizing the Vehicle Routing Problem with Simultaneous Pickup and Delivery and Occasional Drivers (VRPSPDOD). The algorithm is suitable for fleets with both regular and on-demand drivers since it has been especially designed to save overall travel costs and react to changes in demand. The algorithm resolves local optima and produces near-optimal solutions by combining SA for probabilistic exploration with ILS for incremental solution refining.

In a case study involving Soyar Logistics, a Fes, Morocco-based e-commerce logistics company, the algorithm effectively distributed routes to meet different demands. As regular drivers maintained shorter, more reliable routes and occasional drivers took on more complicated routes, fleet efficiency was maximized and the company was able to meet peak demand without expanding its core fleet.

According to the results, the hybrid ILS-SA algorithm significantly decreased travel costs and average route lengths for regular drivers, surpassing traditional heuristics. This adaptive allocation technique highlights the algorithm's scalability and durability in mixed-fleet logistics operations, making it a suitable option for VRPSPDOD in environments that need flexible, cost-effective routing solutions. In dynamic logistics situations, the ILS-SA hybrid strategy shows great promise for improving operational flexibility, route effectiveness, and cost planning.