摘要
近年來，因為高科技產業持續發展，製程技術不斷進步，半導體廠內所用到的新製程和新化學品，相對更具複雜及危險，所以廠內的安全衛生作業管理及異常事故緊急應變處理，需具有更多的人力和物力來維護廠區安全，減少異常災害事故發生之機率，故於廠內編制緊急應變組織(Emergency Response Team，ERT)實施災害事故處理、搶救等相關作業。在廠內的ERT編制中，除了維護廠區安全及杜絕異常危害事故發生外，ERT人員出勤次數多寡也將間接影響人員工作效率及維護廠區安全的救災作業。目前ERT人員輪班是以三班制輪值，加上半導體廠區工作特性為每天24小時持續運作生產，在人力指派上，大多由管理者依經驗判斷，實施人工排班作業，此方法通常較不具效益且未能達公平性之原則，作業上也會耗費較多時間。 本研究針對上述問題以新竹科學園區某半導體廠房為範例，證明本研究模式合理及實用性，以建構一套數學最佳化模型，運用LINGO-9.0套裝軟體實施求解，即可得緊急應變組織ERT人員最平均出勤班次與人力指派最佳化。經研究結果發現本研究模式比人工經驗指派效率來得更佳，且人員排班次數也具公平性，可做為高科技廠房對於緊急應變組織ERT人員人力指派實務上之參考。
並列摘要
In recent years, with the continuous development of high-tech industry and constant advancement of process technologies, the new processes and new chemicals used in the semi-conductor plants are getting more complicated and dangerous. Thus, the safety and health operation management and emergency response of the unexpected accident in the plant require more manpower and resource to guarantee the safety in the plant area and reduce the probability of the unexpected disaster and accident occurrence. That is why the Emergency Response Team (ERT) is set up in the plant to respond to the disaster and accident, and perform first-aid and the related operations. Besides guaranteeing the plant safety and preventing the unexpected disasters and accidents, the attendance of ERT members will also indirectly affect their working efficiency and disaster recovery operation in the safety maintenance. Currently the ERT members work in three shifts system. Moreover, the semi-conductor plants are mostly in 24-hour operation. In terms of the manpower assignment, the managers mostly make decision based on the experience, and carry out manual scheduling operation. This deficient method fails to achieve the fair principle, and consumes much time in operation. For above issues, the study takes a semi-conductor plant located in Hsinchu Science Park as an example, by which the model developed in the study is proven to be reasonable and practical. It constructs a mathematic optimization model and calculates solution by using LINGO-9.0 software package. In this way, it works out the most even attendance schedule and manpower assignment optimization for the ERT members. The study results show the model developed in the study is more efficient than the manual assignment based on experience, which also achieves the fairness in the attendance schedule. It can provide reference for the high-tech plants in the practice of ERT manpower assignment.