摘要
无委托板坯是指炼钢工序剩余的暂时没有合同对象的板坯.无委托板坯匹配问题是研究如何将这些板坯合理分配给热轧计划中的合同.针对实际问题,建立了多目标优化的0-1整数规划模型.鉴于其NP-hard特性,采用蚁群算法(Ant colony optimization,ACO)获得近似解.根据问题特点,提出钢级分解策略,并加入随机扰动策略,构造了合成邻域以改进算法性能.目前,以该算法为核心的决策支持系统已在企业通过应用验证,与人工匹配相比,日匹配板坯量平均提高了52.42%,百吨板坯匹配切损量平均降低了11.36%.
The slabs left in steel-making procedure and not belonging to any order at the moment are called open-order slabs. The open-order slab matching problem is how to assign these slabs to compatible orders in the hot-rolling plan. For an actual problem, a multiple objective 0-1 integer programming model is established. It is an NP-hard problem, so we take the algorithm of ant colony optimization （ACO） to obtain its approximate solution. We propose decomposition strategy based steel grade, meanwhile we construct a random perturbation strategy and the compound neighborhood to improve the performance of the algorithm. At present, the decision support system for open-order slab matching is already verified by application. Compared with the manual work, the system increases the weight of the slabs matched by 52.42 % per day on average and reduces the scraps of 100 ton matched slabs by 11.36 % on average.
出处
《自动化学报》
EI
CSCD
北大核心
2009年第2期186-192,共7页
Acta Automatica Sinica
基金
国家杰出青年科学基金(70425003)
国家高技术研究发展计划(863计划)(2006AA04Z174)
国家自然科学基金(60674084)资助~~
关键词
钢铁计划
无委托板坯匹配
建模
蚁群算法
Steel-planning, open-order slab matching, modelling, ant colony optimization （ACO）
