系统工程理论与实践
繫統工程理論與實踐
계통공정이론여실천
Systems Engineering—Theory & Practice
2011年
7期
1372~1377
,共null页
半导体制造系统 Petri网模型 遗传算法 分时段优化调度
半導體製造繫統 Petri網模型 遺傳算法 分時段優化調度
반도체제조계통 Petri망모형 유전산법 분시단우화조도
semiconductor manufacturing system; Petri net model; GA; multi~phase optimal scheduling
为了更有效优化半导体制造系统生产资源调度,提出了基于分层着色时间Petri网模型的分时段优化调度方案.通过增强基本Petri网的描述能力和引入分层Petri网的思想,根据半导体制造系统的结构建立其改进Petri网模型一分层着色时间Petri网模型,并在模型中引入调度库所将调度算法嵌入到模型中,该模型能有效地克服基本Petri网模型规模膨胀的缺陷;同时将每一生产周期分成若干时间段,利用遗传算法来寻找各时间段内调度规则组合来优化半导体制造系统的动态性能,实现分时段优化调度,编码时,根据机器组的利用率来选择部分机器组为瓶颈机器组,重点调度瓶颈机器组,提高算法的搜索效率;仿真实验结果验证了该调度方案的优越性.
為瞭更有效優化半導體製造繫統生產資源調度,提齣瞭基于分層著色時間Petri網模型的分時段優化調度方案.通過增彊基本Petri網的描述能力和引入分層Petri網的思想,根據半導體製造繫統的結構建立其改進Petri網模型一分層著色時間Petri網模型,併在模型中引入調度庫所將調度算法嵌入到模型中,該模型能有效地剋服基本Petri網模型規模膨脹的缺陷;同時將每一生產週期分成若榦時間段,利用遺傳算法來尋找各時間段內調度規則組閤來優化半導體製造繫統的動態性能,實現分時段優化調度,編碼時,根據機器組的利用率來選擇部分機器組為瓶頸機器組,重點調度瓶頸機器組,提高算法的搜索效率;倣真實驗結果驗證瞭該調度方案的優越性.
위료경유효우화반도체제조계통생산자원조도,제출료기우분층착색시간Petri망모형적분시단우화조도방안.통과증강기본Petri망적묘술능력화인입분층Petri망적사상,근거반도체제조계통적결구건립기개진Petri망모형일분층착색시간Petri망모형,병재모형중인입조도고소장조도산법감입도모형중,해모형능유효지극복기본Petri망모형규모팽창적결함;동시장매일생산주기분성약간시간단,이용유전산법래심조각시간단내조도규칙조합래우화반도체제조계통적동태성능,실현분시단우화조도,편마시,근거궤기조적이용솔래선택부분궤기조위병경궤기조,중점조도병경궤기조,제고산법적수색효솔;방진실험결과험증료해조도방안적우월성.
In order to optimize the resources scheduling of semiconductor manufacturing system (SMS) more effectively, the hierarchical colored-timed Petri net (HCTPN) model based multi-phase optimal scheduling (MPOS) approach was proposed. Through the description enhancement and hierarchical idea for the basic Petri net, an improved Petri net model named HCTPN model was established, the HCTPN model can overcome the risk of model explosion of basic Petri net, and the scheduling place was embedded into the HCTPN model for the MPOS. At the same time, the MPOS was realized by dividing a production cycle into multi phases, the genetic algorithm (GA) was used to optimize the combination of integrated rules at each phase~ the bottleneck machine groups were identified by their utilization rates, and the search efficiency of MPOS was improved by emphasizing the bottleneck machine groups through coding phase of GA. The results of simulation experiments verify the feasibility of the proposed strategy.