How to use it
- 1. Drag customersGrab any red dot and move it; the truck tour and the drone sortie re-solve as you drag.
- 2. Tune the droneRange caps which sorties are feasible; speed ratio is how much faster it flies than the truck drives.
- 3. Read the planNavy = truck route; dashed red arc = the drone's launch → serve → rejoin sortie. ✈ marks the drone-served customer.
- 4. Check the sensitivityThe chart sweeps drone range and plots the saving - find where extra range stops paying.
The planner
Drone parameters
Distances are Euclidean; 10 px = 1 km. The truck drives at 1 km/min, the drone flies faster by the speed ratio.
Saving vs drone range
Completion-time saving of the best truck+drone plan over truck-only, swept across drone range. Your current range is marked.
Try these experiments
- Drag one customer far into a corner - the drone should claim it and the saving jumps (the truck skips the longest detour).
- Cluster all customers tightly around the depot and find a layout where the drone helps nothing - there is no detour worth cutting.
- Drop the speed ratio to 1× and watch when the drone stops being useful - then find the ratio where it starts paying again.
About this tool
The concept
The Flying Sidekick TSP pairs a truck with a drone that serves one customer per sortie - the classic Murray & Chu (2015) model behind Amazon-style hybrid delivery.
How it works
The truck tour is the exact optimal TSP, computed by Held–Karp dynamic programming. For each candidate customer we then enumerate every (launch, rejoin) pair, respect drone range, and keep the completion-time-minimising plan - so the whole plan is genuinely optimal at this size.
Winning strategy
Drones earn their keep on far, isolated customers; in dense clusters the truck alone wins. Range and speed set how far that frontier reaches.