國立臺灣大學電機工程學系暨研究所張時中2006-07-252018-07-062006-07-252018-07-062005-07-31http://ntur.lib.ntu.edu.tw//handle/246246/8030本計畫三年的目標﹕(1) 為排序佳化模擬設計削減 搜尋空間的新方法，(2) 將方法開發成工具模組，作為 整合系統佳化平台的一部份，(3) 將模擬進行排序佳化 應用在次世代半導體廠的有效生產排程。第一年中我們 以平穩(Stationary)馬可夫決策問題為載具，提出模擬排 序策略疊代(OOBPI) 的構想，利用策略疊代的架構，以 模擬排序佳化來估算每一狀態的cost-to-go 函數值與最 佳決策。初步模擬結果顯示這個構想的計算效能較傳統 模擬為基礎的策略疊代法可快十倍。 第二年中我們完成OOBPI演算法的設計，包括方法 理論、收斂分析、數值驗證及推展至一般動態規化的初 探。在半導體晶圓生產模擬器方面，我們加入產品優先 序之考量，以模擬半導體廠具有不同優先次序之生產排 程。目前正結合OOBPI方法與半導體晶圓生產模擬器進 行生產排程組合選擇的研究。The three-year research objectives of this project are (1) to design search space reduction methods for simulationbased ordinal optimization (OO), (2) to develop these methods into tool modules as part of an integrated system optimization platform, and (3) to apply simulation-based OO to effective production scheduling of 300mm foundry fabs. In the first year, we considered the class of Stationary Markov decision problems as the conveyer problem. We proposed an idea of OO-Based Policy Iteration (OOBPI) to handle the combinatorial complexity of decisions over the time axis. Utilizing the framework of policy iteration, we approximate the optimal cost-to-go and optimal decision of each state by simulation-based OO. Preliminary numerical studies indicated one order of speed-up of OOBPI over the traditional simulation-based policy iteration. In the second year, we have completed the design of OOBPI algorithm, including the theories for the method, convergence analysis, simulation study and exploration of possible extension to general dynamic programming. In developing a simulator of semiconductor wafer fabrication with differentiated services, we have incorporated priority service discipline into the simulator. We are now combining OOBPI with the fab simulator to study dynamic composition of production schedules for fabs.application/pdf278579 bytesapplication/pdfzh-TW國立臺灣大學電機工程學系暨研究所排序佳化策略疊代演算法設計生產模 擬產品優先序生產排程Ordinal OptimizationPolicy IterationAlgorithm designFab SimulatorPriorityProduction Schedulingreporthttp://ntur.lib.ntu.edu.tw/bitstream/246246/8030/1/932213E002043.pdf