The Vehicle Routing Problem with Movement Synchronization (VRPMV / VRPMS) extends classic delivery routing by coordinating two complementary resources:
- Heavy Resource (Truck): High capacity and long range, but slower speed and higher operating cost per mile.
- Light Resource (Drone or Courier): Low capacity and battery-constrained range, but faster speed and fractional operating cost.
Rather than returning independently to the depot, the light resource launches from the heavy truck at an optimal customer node, delivers to a high-density customer cluster, and rejoins (rendezvous with) the heavy truck at a downstream customer node.
This two-stage Genetic Algorithm decouples the combinatorial complexity into:
- Stage 1: Solves the macro vehicle routing with time windows using partially matched crossover (PMX) and elitist selection.
- Stage 2: Greedily clusters interior customers within drone capacity/range constraints, and optimizes launch/rendezvous points via custom rendezvous crossover and mutation operators.