The Hot Strip Mill Scheduling Problem With Uncertainty: Robust Optimization Models and Solution Approaches.

Area of Science:

• Operations Research
• Industrial Engineering
• Materials Science

Background:

• Steel industry faces complex hot strip mill (HSM) scheduling challenges.
• Uncertainty in processing times makes deterministic models inadequate.
• Balancing penalty costs and energy consumption (slab reheating) is crucial.

Purpose of the Study:

• To develop robust optimization models for biobjective HSM scheduling under uncertainty.
• To minimize both production penalties and energy consumption for slab reheating.
• To provide effective algorithms for solving uncertain scheduling problems.

Main Methods:

• Formulated a robust optimization (RO) model and an equivalent mixed-integer linear programming (MILP) model.
• Developed an improved Benders' decomposition algorithm for exact solutions.
• Proposed a multiobjective particle swarm optimization (MOPSO) algorithm with local search for large instances.

Main Results:

• The exact algorithm effectively solves smaller HSM scheduling instances.
• The MOPSO metaheuristic achieves satisfactory solution quality for both small and large instances.
• Robust optimization successfully addresses uncertainty in rolling processing times.

Conclusions:

• Robust optimization provides a viable approach for uncertain HSM scheduling.
• Both exact and metaheuristic algorithms demonstrate effectiveness for different problem sizes.
• The proposed methods contribute to energy reduction and cost efficiency in steel production.