A Weak-Robust Modeling and Knowledge-Driven Solution Approach for Scheduling Margin Calculation of Hot Rolling...

Area of Science:

• Industrial Engineering
• Operations Research
• Materials Science

Background:

• Production process uncertainty significantly impacts hot rolling scheduling plan execution.
• Existing robust optimization methods struggle to quantify long-term production capacity (time margin) and account for planning delays and upstream/downstream constraints.
• Accurate scheduling margin calculation is crucial for efficient hot rolling operations.

Purpose of the Study:

• To propose a weak-robust modeling and knowledge-driven solution framework for scheduling margin calculation (WRoKS-SMC).
• To address limitations in quantifying production capacity and incorporating diverse uncertainties in hot rolling scheduling.
• To enhance the robustness and efficiency of hot rolling production scheduling plans.

Main Methods:

• Reformulated the weak-robust optimization problem into a discretized Markov decision process (MDP) model to handle continuous uncertainty.
• Developed a multihierarchical graph convolutional network (MH-GCN) to encode multidimensional slab features for MDP state representation.
• Implemented a knowledge-driven imitation learning framework (KDIL) to integrate hard production scheduling constraints into the MDP.

Main Results:

• The proposed WRoKS-SMC framework effectively calculates scheduling margins under uncertainty.
• Experimental results using practical steel enterprise data demonstrate superior performance compared to state-of-the-art methods.
• The method shows significant improvements in scheduling costs and margin calculation accuracy.

Conclusions:

• The WRoKS-SMC framework provides an effective solution for calculating scheduling margins in uncertain production environments.
• The integration of MDP, MH-GCN, and KDIL offers a robust approach to complex hot rolling scheduling challenges.
• This research contributes to optimizing production planning and capacity management in the steel industry.