AGV fleet put to the test in a stochastic model before it was sized
Give the client an answer it could believe: not just a fleet size, but how often the system would fail with its own data.
- Client
- Confidential client
- Sector
- Sector: Automotive
- Division
- Division: CYP SIRE
Context
The client needed to know how many AGVs its circuit required and whether the buffer in the layout was big enough. The usual calculation, with averages and a safety margin, gave an answer, but not a convincing one.
Challenge
- A calculation based on averages does not capture the real variability of the operation
- The client did not trust the first calculation
What we did
A stochastic model fed with the client's input data plus random variation.
-
Saturation calculation
A first sizing with the usual method: average times and routes, plus a safety margin.
-
Stochastic model
The plant's input data, with random variation, across a large number of scenarios, to measure how often the system would fail.
Results
The model measured the risk of each configuration with the client's own data before the fleet was sized.
A similar project? Let’s talk.
You hear back from the team that does that work: CYP ENGY, CYP DATA or CYP SIRE.