平行機台一直是近年來在排程問題中一個重要的研究領域之一,但平行機台的排程問題在學術上是屬於困難度極高的組合最佳化問題,除了少數特例外,此類問題均屬於NP-hard問題,需要花費相當多的時間及資源才能求得最佳解,因此本研究期望能在具有順序相依整備時間,工作是有不同到達時間且加工方式是可視為可分段處理的前提條件下,去建構出一個以總絕對延誤時間最小化的不相關平行機台排程模式。 而本研究是嘗試使用具有平行搜尋與能避免落入局部最佳解的基因演算法來求解此不相關平行機台之排程問題,並結合某偏光板廠中的PSA製程之實際案例資料去對在部份分段作業模式下的不同分段比重進行效益比較,期能提供業界或後續研究人員作為參考。
Parallel-machine scheduling has been one of the important research fields of scheduling problems in recent years. It is a difficult combinatorial optimization problem. Except for few cases, this kind of problems belongs to NP-hard problems which require considerable time and resources to find optimal solutions. This research considers parallel-machine scheduling problems where jobs can be divided into stages and do not require continuous process. An unrelated parallel-machine scheduling model with dividable jobs, sequence-dependence setup times, and non-zero arrival times, is constructed and the objective is to minimize total absolute lateness. This research applies Genetic Algorithm Approach on the considered scheduling problem. The Genetic Algorithm has parallel searching functions and capability to avoid partial optimal solution. Numerical experiment containing various test problems with real case data of the PSA process from a Polarizer manufacturing factory is conducted to evaluate the performance of the proposed algorithm. It is the aim of this research that the results can be of valuable to industries and follow-up research.