A Hierarchical Two-Level Adaptive Allocation Framework for Multi-Location Inbound Logistics Under Operational Constraints

Abstract

Background: The allocation of non-divisible inbound deliveries across multiple warehouses requires the simultaneous consideration of capacity constraints, category-specific restrictions, workload balance, and long-term allocation consistency. Methods: This study proposes a hierarchical two-level allocation framework combining strategic category-rotation policies, normalized marginal scoring, and signed historical feedback. The algorithm is executed once per day to generate warehouse assignments for the following operational day. Historical correction is based on a rolling window covering the preceding 30 daily planning periods. Results: The framework was evaluated using daily simulation instances ranging from 100 to 1500 pallets, with an average of approximately 130 lots per pallet. Across all evaluated instances, the complete allocation procedure was completed in less than 5 s on the specified test system. The results indicate balanced warehouse utilization, progressive reductions in category–location imbalance, stable historical correction, and preservation of hard operational constraints. Conclusions: The framework provides an interpretable and computationally efficient approach for next-day inbound allocation. By combining explicit feasibility filtering, strategic policy signals, and a 30-day historical correction mechanism, it supports both short-term operational decisions and longer-term allocation balance.