CAREER: Dynamic Stochastic Scheduling

9625061 Duenyas本次CAREER奖的研究重点是在存在可变性的情况下动态生产控制的新技术、结构结果和近似。该研究在三个主要领域开发了有效的控制机制：批量处理机器，具有重要设置时间的机器和时间限制（“易腐烂”）处理器。研究者将应用半马尔可夫决策理论、封闭排队网络分析和随机排序的结果来获得小问题的结构结果，这些结果将指导大问题的启发式技术的发展。该教学计划包括开发一门新课程，将制造过程和计划整合在一起，该课程将与联邦政府资助的教育联盟一起提供，该联盟的使命是在弱势城市人口中发展制造专业知识。本研究解决的生产控制问题与半导体设备和平板显示器的制造特别相关，这两个行业都是美国军事和经济竞争力的关键行业。研究者将与这两个行业的制造商密切合作，包括为平板显示行业开发商业工厂计划/调度软件。这项研究有可能提高这些制造设施的产量和吞吐量。

9625061 Duenyas The research focus of this CAREER award is on new techniques, structural results, and approximations for dynamic production control in the presence of variability. The research develops effective control mechanisms in three main areas: batch processing machines, machines with significant setup times, and time-constrained ("perishable") processors. The investigator will apply results from semi-Markov decision theory, closed queueing network analysis, and stochastic ordering to obtain structural results for small problems that guide the development of heuristic techniques for larger problems. The teaching plan includes the development of a new course integrating manufacturing processes and planning that will be offered in conjunction with a federally-funded educational coalition whose mission is the development of manufacturing expertise in a disadvantaged, inner-city population. Production control problems addressed in this research are particularly relevant in the manufacture of semiconductor devices and flat-panel displays, both critical industries for U.S. military and economic competitiveness. The investigator will work closely with manufacturers in both industries, including a developer of commercial factory planning/scheduling software for the flat-panel display industry. This research has the potential to increase both yield and throughput in these manufacturing facilities.