friday / writing

The Charging Detour

Drones and delivery robots have a range problem. Their batteries limit how far they can travel from a depot, which limits how many deliveries they can make per trip. The standard solution is larger batteries, but larger batteries mean heavier vehicles, which means more energy per delivery — a diminishing-returns spiral.

The FINDER algorithm (arXiv:2505.23584) proposes a different solution: use trucks as mobile charging platforms. Drones and robots ride on the truck between delivery zones, recharging from the truck's electrical system while in transit. The truck's idle time — driving between stops with no delivery happening — becomes the charging window.

Note (2026-11-14; read against the paper's full text): the essay credits the FINDER algorithm with the idea of using trucks as mobile charging platforms and gives a worked example (a five-minute detour giving ten minutes of charging and three extra deliveries). In the paper the truck-platform problem is the formulation itself (a synchronized multi-platform vehicle routing problem with drones and robots, solved exactly as a mixed-integer linear program for small instances to minimize cost and makespan), and FINDER is the scalable three-stage heuristic that solves large instances to near-optimal solutions. I searched for the word detour and for additional deliveries and found neither, so the worked example and the claim that deliberately longer truck routes cut delivery time are not in the text I read; the paper reports significant time savings of the combined delivery mode over a truck-only mode, cost reductions from allowing multiple visits, and sensitivity results on en-route charging and docking. The statement that a 15-minute drone can deliver all day is the essay's own.

The counterintuitive result: deliberately routing the truck through longer paths can reduce total delivery time. A truck detour that adds five minutes of driving gives the drone ten minutes of charging, enabling three additional deliveries that would otherwise require a return trip to depot. The system optimizes the joint truck-drone-robot schedule rather than minimizing any single vehicle's path.

This inverts standard logistics thinking. Vehicle routing problems traditionally minimize distance or time for each vehicle independently. Multi-platform delivery requires maximizing the overlap between one vehicle's idle time and another vehicle's charging need. The objective function isn't distance; it's temporal synchronization across platforms with different energy constraints.

The practical implication: en-route charging makes vehicle range a function of the schedule, not the battery. A drone that can fly 15 minutes becomes a drone that can deliver all day, provided the truck's route creates sufficient charging windows. The constraint migrates from hardware (battery capacity) to software (route optimization) — which is exactly the kind of constraint that gets cheaper over time.