# Corrin & Vale Unit 4 — proposal for the Operations Director

Recommend **Option B: move all four packing benches south of the pick zones and use three BF-T1 robots**, with three chargers in the customer's east charging area. Retain outbound staging and the docks. Treat this as the preferred layout for commissioning validation; the simulation does not establish a guaranteed 110-tote/hour service level or prove that three is the minimum fleet.

## Evidence

The saved A/B study compares four robots at the existing benches (A) against three at the final south benches (B), using 20 matched seeds, a 30-minute peak-demand preload and the full 2.5-hour peak observation. Native completion rates average **102.38 jobs/hour for A and 112.06 for B**; the paired improvement is **9.68/hour (95% confidence interval 7.30–12.06)**. These counts include charging jobs. A is below the 110-tote/hour demand even before those jobs are removed.

For the first seed, packed-tote output is **99.2/hour for A and 108.8/hour for B**. Mean P&D wait is **8.84 minutes versus 1.05 minutes**. These are single-seed measurements, not 20-seed estimates. Across all 20 seeds, robot utilisation averages **98.69% for A and 85.69% for B**, including charging and loaded waiting.

An earlier south layout, with slightly different bench positions, has a full 20-seed tote audit: **108.96 packed totes/hour (95% interval 106.46–111.46)** and **2.34 minutes mean P&D wait**. Its existing-bench three-robot counterpart produced 94.22/hour and a 15.30-minute mean wait. Keep this earlier evidence separate from final B. The CSV also retains the earlier four-robot screen. No customer wait SLA was supplied; confidence intervals describe mean simulation output, not a service guarantee.

## Inputs and assumptions

The customer brief specifies 110 orders/hour, one tote per order, zone shares A/B/C of 45/35/20%, four fixed packers and 100 seconds average packing time, mostly around 90 seconds and ranging from 60 to 180 seconds. Picking and the 2.6 lines/order are outside the transport model. Finished-tote arrivals are assumed independent Poisson streams. Packing uses a 75% triangular(60,90,120 s) / 25% triangular(90,120,180 s) mixture, giving a 100-second mean.

The BF-T1 datasheet supplies one tote per robot, 1.0 m/s study speed, a 0.5 m routing radius, 12-second pickup and unload, nearest-free dispatch, idle waiting at the last drop, six minutes charging per hour of work, and one 0.6 × 0.4 m charger per robot with 1.5 m clear in front. Bench capacity is assumed one, reserved before travel through packing, with no extra drop buffer. Charging is approximated by staggered 360-second fleet-clock jobs. Saved A has four robots but only the same three charging streams as B; the earlier four-robot screen includes a fourth stream.

Final B benches have top-left coordinates **(6,27), (13,27), (20,27), (27,27) m**. Chargers 1–3 face west on an assumed mounting rail supplied from east-wall power. The fourth charger shown is optional and unused by B. Rail/power provision and all drawing-derived dimensions require site confirmation.

## Limits and handover

The model omits robot collisions, avoidance, congestion slowing, people interactions, turning orientation, physical tote queues/picker blocking, outbound handling after packing, failures and staff breaks. Charging has no individual battery/SOC model, energy guarantee, docking model or charging fairness. Verify the datasheet's 1.2 m one-way / 2.0 m two-way lane clearances and site safety arrangements separately.

The PDFs show final B; the route sheet contains the first seed's actual complete robot traces, including warm-up. The inherited floor title still says “Rev C existing”; the plan revision identifies Option B. Metric definitions and missing-interval explanations are in `CV-fleet-metric-notes.md`. The full project is exported and saved locally in the browser as **Corrin & Vale Unit 4 — Proposal — Option B**, revision 2; the active B result is current. Customer and robot are fictional.