系统工程学报
繫統工程學報
계통공정학보
JOURNAL OF SYSTEMS ENGINEERING
2001年
2期
121-127
,共7页
钢铁工业%库存管理%板坯倒垛%最优化%启发式算法
鋼鐵工業%庫存管理%闆坯倒垛%最優化%啟髮式算法
강철공업%고존관리%판배도타%최우화%계발식산법
板坯库用于存放由炼钢-连铸工艺阶段产生的、供热轧生产阶段进一步轧制成带钢或钢板的 板 坯.一个板坯库中可分为多个逻辑区域,每个逻辑区域又进一步分为多个垛位.同一垛位的板 坯之间是叠放的.最优倒垛问题是如何从M个相关的板坯垛位中选择N个板坯,在满足 相应的 轧制单元计划所需要的板坯序列要求的条件下,保证总的板坯倒垛次数最少.首先对倒垛问题 进行了定义,在分析了问题的邻域结构基础上,构造了新的多回路启发式算法.通过精选的随 机产生的200个问题例子的实验显示,提出的启发算法的性能在计算速度和最优性方面都明显 优于原系统的启发式算法,在最优性方面平均改进率达到20%,在计算时间方面,可降低一半以 上.新的多回路启发式算法不但提高了车间作业率,也大大降低了生产过程控制机(FLS)的负荷.
闆坯庫用于存放由煉鋼-連鑄工藝階段產生的、供熱軋生產階段進一步軋製成帶鋼或鋼闆的 闆 坯.一箇闆坯庫中可分為多箇邏輯區域,每箇邏輯區域又進一步分為多箇垛位.同一垛位的闆 坯之間是疊放的.最優倒垛問題是如何從M箇相關的闆坯垛位中選擇N箇闆坯,在滿足 相應的 軋製單元計劃所需要的闆坯序列要求的條件下,保證總的闆坯倒垛次數最少.首先對倒垛問題 進行瞭定義,在分析瞭問題的鄰域結構基礎上,構造瞭新的多迴路啟髮式算法.通過精選的隨 機產生的200箇問題例子的實驗顯示,提齣的啟髮算法的性能在計算速度和最優性方麵都明顯 優于原繫統的啟髮式算法,在最優性方麵平均改進率達到20%,在計算時間方麵,可降低一半以 上.新的多迴路啟髮式算法不但提高瞭車間作業率,也大大降低瞭生產過程控製機(FLS)的負荷.
판배고용우존방유련강-련주공예계단산생적、공열알생산계단진일보알제성대강혹강판적 판 배.일개판배고중가분위다개라집구역,매개라집구역우진일보분위다개타위.동일타위적판 배지간시첩방적.최우도타문제시여하종M개상관적판배타위중선택N개판배,재만족 상응적 알제단원계화소수요적판배서렬요구적조건하,보증총적판배도타차수최소.수선대도타문제 진행료정의,재분석료문제적린역결구기출상,구조료신적다회로계발식산법.통과정선적수 궤산생적200개문제례자적실험현시,제출적계발산법적성능재계산속도화최우성방면도명현 우우원계통적계발식산법,재최우성방면평균개진솔체도20%,재계산시간방면,가강저일반이 상.신적다회로계발식산법불단제고료차간작업솔,야대대강저료생산과정공제궤(FLS)적부하.
The slab yard can be used to store slabs produced in steelmaking-continuous cas ting process and required to further form strips and plates by the hot r olling mill. A slab yard can be divided into a large number of logic sections e ach including many pile locations. Slabs are stacked in the slab yard according to the pile location. The optimal turned-out slab pile (TOSP) problem is to d etermine N slabs from M related pile locations so that the total turned -out slab pile number can be minimized while the corresponding slab sequence re quirements of the rolling turn are satisfied. This paper first gives the definit ion of the optimal TOSP problem. Then, a new multiple loop heuristic algorithm i s constructed and implemented. The computational experiments with 200 selected p roblem instances produced randomly show that the new proposed heuristic algorith m performs remarkably better than the old one. The average optimality of the alg orithm is enhanced by 20 % and the computation time is reduced by about 50 %. Th e new algorithm not only increases the shop floor productivity, but also cuts do wn greatly the workload of FLS.