Hi — I built Wessim, a free warehouse throughput simulator that runs in the browser. You build a warehouse operation across 14 stages dock to dock, with 46 equipment types — AS/RS, AMRs, shuttles, sorters, conveyors — and simulate the flow. It shows the binding constraint and what late orders cost against SLA and carrier cutoffs. No sign-up, nothing uploaded.
I spent years doing WMS implementations before I built this, so the warehouse side I know. The part I want feedback on is the rendering side.
The sim runs on a discrete-time fixed one-second clock — not discrete-event — and the 3D floor is rendered from the sim state. Every order, every piece of equipment, every queue on the floor has to match the numbers, or the bottleneck the picture shows you is a lie. Keeping the render honest while the sim runs — that sync between the fixed-clock engine and the 3D scene — was the hard part of the build.
I checked the queueing engine against Erlang C on one test — three pack stations, 1,200 hours each, same station, same random demand, no breaks or faults. Wait time within 1–3%, utilization within 0.5%.
I’d like a second pair of eyes on the rendering and the sync: what would you check first, and what breaks first when a sim and a scene have to agree every tick?
A 30-second clip of an order moving through it: https://youtu.be/1nd39oZmtBI