Reward-Density Heuristic for Dynamic Multi-Vehicle Routing: Performance and Computational Efficiency