optibox
a greedy routing and dispatch system for a multi-shuttle warehouse, built at hackupc 2026 in barcelona (europe's biggest student hackathon) for the inditex challenge - it receives boxes, stores them, ships them to active pallets, and replays the whole run visually.
the problem
inditex's challenge was a warehouse where shuttles move boxes between inbound, storage silos and outbound pallets. the goal was to receive, store and deliver boxes as fast as possible while balancing route efficiency, storage placement, and which pallets are active - all at once.
design choices
- a greedy rule set built around one principle: deliver and store on the same route, so shuttle trips do double duty.
- inbound boxes go either straight to cross-dock delivery or into storage with pickup on the same route.
- a deterministic simulation of 32 shuttles across a silo layout, so any run can be replayed and debugged exactly.
- python + flask backend for the simulation state machine, shuttle task logic, dispatch and pallet lifecycle rules, csv state generation and timeline export; react + typescript + vite + tailwind frontend for playback controls, the silo grid, an event log and a pallet panel.
challenges
- balancing route efficiency, storage placement, and pallet activation/completion at once.
- keeping the simulation deterministic while still being interactive.
- translating backend state transitions into smooth visual playback.
- keeping frontend and backend payload contracts aligned as features evolved.
what i learned
how to manage complex warehouse constraints with clear state transitions, how much data contracts matter when playback depends on backend tick data, and how to collaborate quickly under hackathon time pressure.
team
built with olivia wong, ioannis pialopoulos and angelos skliros.