Seasonal Booking Workflow

Every November, the distribution center in Ipoh faces the same problem: inbound pallets arrive in waves, but the booking system treats every day like a quiet Tuesday. The result is a queue of AGVs waiting at the staging area while the floor crew manually re-prioritizes jobs. This project replaced that reactive process with a seasonal booking workflow that matches robot capacity to actual inbound volume.

The core constraint was simple: no new hardware. The existing fleet of twelve heavy-duty AGVs had to handle a 40 percent spike in pallet volume without slowing down the outbound line. Instead of adding robots, we changed how the fleet receives work. The booking workflow now accepts inbound slots up to 72 hours ahead, assigns each slot a priority tier, and lets the fleet controller batch similar destinations into single trips.

The tricky part was the tradeoff between flexibility and predictability. Warehouse managers wanted to override the schedule when a truck arrived early, but every override disrupted the optimized route plan. We settled on a two-tier system: hard bookings that lock in 24 hours before arrival, and soft bookings that the controller can reshuffle up to six hours prior. This gave the floor team enough room to react without turning the whole schedule into guesswork.

• Inbound slots are booked by truck arrival window, not by pallet count, so the controller can plan around dock availability.
• Priority tiers separate urgent cross-dock loads from standard storage pallets, preventing low-priority cargo from blocking the main aisle.
• Batch grouping pairs pallets headed to the same rack zone, cutting empty travel distance by roughly 18 percent during peak weeks.
• Override logs feed a monthly review, so the team can see which manual changes actually improved throughput and which ones just added noise.

Implementation took six weeks, with the first two spent on data collection. We tracked every manual override for a month to understand why the old system failed. Most overrides happened because the booking form lacked a field for pallet dimensions, so the controller could not tell whether a load would fit on a single AGV. Adding that field alone removed a third of the manual interventions.

The workflow went live in early October, just before the peak season. The first week was rough: operators forgot to mark soft bookings, and the controller had to chase down missing data. By the third week, the rhythm settled. The fleet handled the November surge with the same twelve robots, and the staging area queue never exceeded four pallets during a shift.