New Heuristic Boosts Dynamic Multi-Vehicle Routing Efficiency.

This study introduces a novel reward-density heuristic, termed the Efficiency heuristic, designed to optimize dynamic multi-vehicle routing problems. These problems, which combine aspects of the Vehicle Routing Problem (VRP) and the Orienteering Problem (OP), require a fleet of vehicles to maximize collected rewards within a fixed time horizon while continuously replanning as new tasks emerge. The research evaluates this heuristic across two practical application domains: autonomous drone task allocation and urban taxi dispatch, considering various fleet sizes and task scales. The Efficiency heuristic was benchmarked against four classical construction heuristics and three sophisticated metaheuristic algorithms, including Adaptive Large Neighbourhood Search, Genetic Algorithm, and Simulated Annealing, all under identical testing conditions. The results consistently showed that the proposed Efficiency heuristic achieved solution quality comparable to the best metaheuristic algorithms. Crucially, the Efficiency heuristic accomplished this while requiring two to three orders of magnitude less computational planning time. This establishes its Pareto dominance over all competing methods in terms of the reward-versus-compute frontier, making it exceptionally well-suited for real-time, online deployment in dynamic, time-constrained routing environments. The findings suggest that well-designed greedy heuristics can be superior to complex search procedures for practical applications.