An Improved Self-adaptive Genetic Algorithm for Scheduling Steel-Making Continuous Casting Production

Ling Li,Qiuhua Tang,Peng Zheng,Liping Zhang,C. A. Floudas
DOI: https://doi.org/10.2991/978-94-6239-148-2_40
2015-10-13
Abstract:The steel-making continuous casting scheduling problem holds more constraints than flow shop in no-idle machine constraint, multiple buffers and batch scheduling, resulting in strong NP-hardness. Its mathematical programming model is first established based on unit-specific event-point continuous-time representation. Then a novel improved self-adaptive genetic algorithm (SAGA) is proposed to optimize the sequence among casts with the objective of reducing the total idle times on all machines and minimizing the make-span. In SAGA, the probabilities of crossover and mutation rate are rectified exquisitely and automatically so as to avoid being trapped in local optima and neighbourhood-based mutation operation is adopted to improve the diversity. Experimental comparisons with GAMS/CPLEX and other two state-of-art algorithms demonstrate the effectiveness and efficiency of SAGA in solving the large-size problems.
What problem does this paper attempt to address?