This simulation project modeled a multi-deep AMR-based pallet storage and retrieval system. The goal was to assess system throughput performance under realistic operating conditions, specifically, the digging behavior where retrieving a pallet in the back of a bay requires swapping or relocating blocked pallets.

Key Objectives:
System Modeling — develop a fast, streamlined model in FlexSim to replicate the core behavior of the AS/RS.

Multi-Deep Storage Logic — simulate digging operations where front-position pallets must be relocated to access those in the back, then model the impact of placing these pallets in random or nearby empty slots rather than returning them to their original locations.

SKU Heatmap Visualization — color-code SKUs by velocity: red (highest volume), green, blue, yellow, and orange (lowest volume), to visualize slotting outcomes.

Throughput Validation — assess whether the system could meet throughput requirements despite the randomized swap strategy and robot congestion.

Slotting Pattern Evaluation — explore whether allowing swapped pallets to “settle” naturally leads to high-velocity SKUs (red) migrating to front positions over time.

Results:
The simulation provided a rapid, low-complexity method to validate system capacity and explore the operational trade-offs of multi-deep storage strategies. It also highlighted that while random swapping does not inherently lead to optimal slotting, the overall system was robust enough to meet throughput requirements without precise slotting control and despite some robot congestion.