智慧型基因演算法於流程型工廠排程之應用─以太陽能板為例

Using Intelligence Genetic Algorithm for Flowshop Scheduling Problems-A Case Study of Solar Cell Plant

洪振家(Chen-Chia hung)

指導教授 ： 張百棧

元智大學/工程學院/工業工程與管理學系/碩士(2008年)

摘要

近年來，伴隨著科技的發展，人類物質享受不斷的提升，但天然資源卻是越來越少，天然氣、石油的蘊藏量逐漸的減少，科學家預言地球的石油產量只夠人類再使用五十年，在這迫在眉睫的時刻裡，替代能源的開發是不容怠慢的，太陽能電池便是最潔淨的替代能源之ㄧ。雖然工業技術越來越發達，但大量生產卻不再是工業時代的重點，生產不只是要大量生產，更是需要講求效率的提升與成本的控制，近幾年來，企業注重的重點例如總完工時間（total completion time）、最大延遲時間（maximum tardiness）、總延遲時間（total tardiness）、最大流程時間（maximum flow time）等都是企業所注重的生產衡量績效。因此，生產排程的規劃工作就顯得格外重要。
現場排程在實務作業裡總會有些問題在上列各項衡量準則中往往是衝突的，我們不能只一昧的追求其中一項之最佳化，因為很有可能會造成另一個績效衡量準則的損失，這也是為什麼多目標排程最佳化的問題會逐漸被大家所重視的
原因。本研究利用智慧型基因演算法來求解流程型機台的問題，以總完工時間、總延遲時間最小化為排程目標，並實際運用在太陽能板產業上，利用基因演算法獨特搜尋特性，求得符合績效的最佳排程，比較企業排程的時間，再從彼此時間的差異性來轉換成實際產能，以得到理論與實務的真正結合。

關鍵字

智慧型基因演算法 ； 流程型機台 ； 實際產能

並列摘要

In the recent years, along with the improving technology, the quality of modern living is getting better and better. But the nature resources is getting less and less. The quantity of gas and fuel is reducing. The scientists oversee that fuel is going to run out in 50 years or so. In this critical moment, finding another kind of substitute energy is strongly urgent. Solar battery is one of the most clean substitute energy. Although the technology of the industries is becoming mature, the massive production is no longer taking the lead in the hi-tech society. Focusing the promoting effect and controlling the cost while making the massive production are more important than ever. During the recent years, total completion time, maximum tardiness, total tardiness and maximum flow time are the major effects that have been measured during the production process in the enterprise. So to speak, the scheduling of the production planning is extremely critical.
In the realistic manufacture, we may foresee some conflicts happen in the shop floor scheduling while trying to meet the standard of which mentioning above. We can’t just pursue the optimistic of one method, and fail to measure another method. This is exactly the reason why people value the most the optimistic of multi-purposes of the production planning. The research is about using intelligent genetic algorithm to solve the problem of the workflow machines, setting the planning goal in minimum the total completion time and the total tardiness, and practicing practically in solar modules industry. Using the unique searching specialty of genetic algorithm locates the best production planning which meets the effects of all. Compare other production planning of the enterprises and form the practical productivity through the time gap differences, in order to combine the theory and practical matter.

並列關鍵字

Genetic algorithm ； Workflow machines ； Practical productivity

參考文獻

17. 包濬瑋，「太陽光發電系統運轉性能評估」，碩士論文，中原大學，2003。

32. 湯璟聖，「動態彈性平行機群排程的探討」，碩士論文，中原大學， 2003。

27. 陳啟嘉，「基因結構探勘於承接式子群體基因演算法求解多目標組合性問題

23. 林昆霖，「子群體基因演算法於多目標排程之應用－以PCB鑽孔作業為例」，碩士論文，元智大學， 2005。

26. 陳正雄，「塔布搜尋法在塑化業排程之應用-以BOPP FILM 為例」，碩士論文，元智大學，2000。