Precast component production scheduling considering outsourcing and multiskilled resources

Precast component (PC) production is a critical stage within the prefabricated construction supply chain. Motivated by a practical challenge observed in PC manufacturing, this study investigates an integrated PC production scheduling problem that concurrently addresses outsourcing decisions and multiskilled workforce constraints from the manufacturer's perspective. This integration presents significant complexity, requiring the joint optimization of two interdependent subproblems: (1) determining the quantity of PC orders to outsource, and (2) scheduling self-produced orders across production lines while sequencing tasks under multiskilled workforce constraints. To address this challenge, a novel model is established to describe the decision-making problem and is then linearized as an integer linear programming (ILP), which is solvable by the off-the-shelf commercial optimizer for small-scale instances. For computationally intensive large-scale problems, an efficient Deep Q-network (DQN) algorithm is designed. The proposed methods are validated through extensive computational experiments based on a real-world PC production case, including comparative analysis against alternative methods and sensitivity analysis of key factors. Compared with the prevalent empirical approach, the proposed methods yield average monthly cost savings exceeding 31,000 Chinese yuan (CNY) across both shoulder and peak seasons. Furthermore, primary findings and managerial insights are derived from experimental results to assist managers in making decisions concerning outsourcing, self-production scheduling, and the flexible allocation of multiskilled employees.