Hierarchical Optimization for Operationally-Constrained Resource Planning

Effective resource planning presents a broad problem across industries due to inner operational constraints. Existing methods are hard for generalization because they entail either unique models or customized specific solutions. To address this challenge, we pose the Operationally-Constrained Resource Planning Problem (OCRPP), which abstract assets and resources using job concepts to decouple the entangled constraints to describe the resource planning problem in a standard way. Meanwhile, we propose the Hierarchical Allocation Optimizer (HAO), a meta-heuristic solving framework containing 3 phases, to speed up feasible solution search with lower bound estimation. Experiments show HAO’s superiority in terms of time and objectives compare to the alternatives, which adopts rule-based heuristic or general meta-heuristic. HAO’s rapid response and flexibility are demonstrated to show its capability to adapt to various business scenarios and applications such as airlines, logistics companies, and so on.