A novel hybrid algorithm for scheduling steel-making continuous casting production
A novel hybrid algorithm for scheduling steel-making continuous casting production
复制标题
DOI：
10.1016/j.cor.2008.10.010
复制
发表时间：
2009-08-01
影响因子：
4.6
通讯作者：
Tarkesh, Hamed
中科院分区：
文献类型：
--
作者：
Atighehchian, Arezoo;Bijari, Mehdi;Tarkesh, Hamed
研究了炼钢连铸调度问题。该问题是炼钢工艺约束下混合流水车间调度问题的一个特例。由于经典的优化方法无法在适当的时间内得到最优解，因此提出了一种新的迭代算法。该算法，命名为HANO，是基于蚁群优化（ACO）和非线性优化方法的组合。HANO的解决方案构建分为两个阶段。第一阶段确定离散变量（对应于作业机器分配和排序），而第二阶段确定连续变量（对应于其分配的机器上的工件的时间）通过非线性优化方法HANO的效率进行了比较，作为一个真实的的情况下，在Mobarakeh钢铁公司（MSC），中东最大的钢铁厂使用的启发式算法。通过算例分析，将该算法与遗传算法进行了比较，结果表明，该算法比MSC的启发式算法具有更高的效率。最后，将HANO算法与GA算法进行了效率比较，结果表明，HANO算法在95%以上的情况下具有较好的性能，而在剩下的5%的情况下，其性能效率没有差异。(c)2008爱思唯尔有限公司版权所有。
In this paper, steel-making continuous casting (SCC) scheduling problem (SCCSP) is investigated. This problem is a specific case of hybrid flow shop scheduling problem accompanied by technological constraints of steel-making. Since classic optimization methods fail to obtain an optimal solution for this problem over a suitable time, a novel iterative algorithm is developed. The proposed algorithm, named HANO, is based on a combination of ant colony optimization (ACO) and non-linear optimization methods. The solution construction in HANO is broken up into two phases. The first phase determines the discrete variables (corresponding to job-machine assignment and sequencing), while the second phase determines the continuous ones (corresponding to timing of the jobs on their assigned machines) through a non-linear optimization method.The efficiency of HANO is compared with a heuristic algorithm as a real case used at Mobarakeh Steel Company (MSC), the biggest steel factory in the Middle East. In addition, the proposed algorithm is compared with Genetic Algorithm, as a search method for both discrete and continuous variables, through solving several instances.Numerical results reveal the higher efficiency of the proposed approach compared with the heuristic one used at MSC. Furthermore, the efficiency of HANO is compared with GA to show that HANO enjoys a better performance in more than 95% of the cases while in the remaining 5%, its performance efficiency shows no difference. (c) 2008 Elsevier Ltd. All rights reserved.