Abstract: Aiming at the special process requirements of steelmaking-continuous casting scheduling and considering the principle of furnace matching and multiple refining stages,a constraint satisfaction model was established with the objectives of furnace-caster matching degree,the total waiting time and the advance/delay time of the cast.For its multi-objective characteristics,an improved fast elitist Non-Dominated Sorting Genetic Algorithm (NSGA-Ⅱ) algorithm with adaptive grid selection strategy was proposed,which overcame the problem of easily losing the useful information with traditional Pareto domination method.The experiments and compared algorithm were set up based on actual production data of various scales,and the results showed that the proposed improved NSGA-Ⅱ algorithm proposed was superior to the traditional NSGA-Ⅱ in terms of convergence,diversity of optimal solution set and computational efficiency based on the scheduling problems of steelmaking-continuous casting.

Key words: steelmaking-continuous casting, prodution scheduling, multi-objective optimization, adaptive grid algorithm, genetic algorithms, fast elitist non-dominated sorting genetic algorithm

摘要： 针对炼钢连铸调度的特殊工艺要求,在考虑炉机匹配原则和多重精炼的情况下,建立了以炉机匹配度、炉次间等待时间、浇次的开浇提前/拖期时间为评价指标的多目标约束满足优化模型,并针对其多目标特征,提出一种基于自适应网格法的择优策略来改进带精英策略的快速非支配排序遗传算法,有效克服了使用传统Pareto支配法择优策略在解决离散问题时容易丢失有用信息的缺陷。基于多种规模的实际生产数据进行仿真实验,结果表明所提算法在收敛性、最优解集多样性和计算效率方面优于传统带精英策略的快速非支配排序遗传算法。

关键词: 炼钢&mdash, 连铸, 生产调度, 多目标优化, 自适应网格技术, 遗传算法, 带精英策略的快速非支配排序遗传算法