本論文緣起於產學合作太陽能工廠中晶碇到晶片製程之最佳派工研究。其中所採用的績效指標為七天穩態情況下每一元作業員薪水所得到之產品銷售利潤，決策變數為需被服務之工件來到每個製程中被指派機台編號及作業員編號。
我們提出多項式機率分佈派工法則，以「人與工件之距離、作業人員能力與其薪水之間之函數」作為指派的依據。經模擬實驗(Flexsim) 顯示，本研究所提出之派工法則優於傳統“再生之經驗暨啟發式(meta-heursitic)”演算法，如基因演算法(Genetic Algorithm)及粒子群聚最佳化(Particle Swarm Optimization)演算法在合理的搜尋時間(24小時)所得之解。模擬實驗也顯示，本研究建議的派工比目前產學合作工廠所使用的派工法則提升了51% 的績效。

Motivated by the dispatching problems in our collaborated solar company, we investigate the optimal dispatching rule for machines and manpower allocation in the solar ingot to water manufacturing systems. The performance of the method is measured by the sales profit for each dollar paid to the operators in a 7 days time window in steady-state. The decision variables are identification-number of machines and operators when each job is required to be served in each process.
We propose a “Multi-nominal distribution dispatch rule” which is a function of operator’s ability, salary and the corresponding location. Simulation (via Flexsim) results show that the proposed rule outperforms many traditional meta-heuristic algorithms such as Generic algorithm (GA) and Particle Swarm Optimization (PSO). Simulation results also show that the proposed dispatch rule demonstrates improvement on the sales profit metrics of about 51% comparing with the method currently used in our collaborated solar company.

第1章 緒論
1.1 研究背景
1.2 多晶矽太陽能晶碇到晶片製程
1.3 研究動機與目的
1.4 名詞與符號定義
1.5 研究主題與相關課題
1.5.1 系統參數
1.5.2 穩態與暫態
1.5.3 績效評估與討論
1.5.4 決策變數
第2章 文獻回顧與潤飾
2.1 基因演算法(Genetic Algorithm, GA)
2.2 粒子群聚演算法(Particle Swarm Optimization, PSO)
第3章 本研究所提出之派工法則
第4章 模擬模式構建
第5章 績效分析與比較
5.1 整備時間為常數之情況
5.1.1 目前方法(最短距離派工)之績效
5.1.2 本文所提方法(多項式機率派工)之績效
5.2 整備時間為常態分配
5.3 模擬時間探討
第6章 結論與未來研究
6.1 結論
6.2 未來研究
參考文獻
附錄

[1] Agnetis, A., Murgia, G. and Sbrilli, S.(2013)A job shop scheduling problem with human operators in handicraft production. International Journal of Production Research, 52, 13, 3820-3831.
[2] Asefeh, H.-G., Masound, R. and Neda, M.(2012)A novel mathematical model for manpower scheduling in break (relief) times in mixed model assembly lines. Procesia-Social and Behavioral Sciences, 62, 1371-1377.
[3] Chang, A.K., Jiang, H.-Y., Di, Y., Zhu, D. and Ge, Y (2008)Time-line based model for software project scheduling with genetic algorithms. Information and Software Technology, 50, 1142-1154.
[4] Lobo, B.J., Hodgson, T.J., King, R.E. and Thoney, K.A.(2013)Allocating job-shop manpower to minimize Lmax: Optimality criteria, search heuristics, and probabilistic qquality metrics. Computers and Operations Research, 40, 2569-2584.
[5] Mencia, R., Sierra, M.R., Mencia, C. and Varela, R (2015)Memetic algorithms for the job shop scheduling problem with operators. Applied Soft Computing, 34, 94-105.
[6] Pan, Q.-K., Suganthan, P.N., Chua, T.J. and Cai, T.X (2010)Solving manpower scheduling problem in manufacturing using mixed-integer programming with a two-stage heuristic algorithm. The International Journal of Advanced Manufacturing Technology, 46, 1229-1237.
[7] Parisa, S.-S., Reza, T.-M. and Hamed, K.(2013)Solving a new fuzzy multi-objective model for a multi-skilled manpower scheduling problem by particle swarm optimization and elite tabu search. The International Journal of Advanced Manufacturing Technology, 64, 1517-1540.
[8] Yoshimura, M., Fujimi, Y., Izui, K. and Nishiwaki, S (2006)Decision-making support system for human resource allocation in product development projects.International Journal of Production Research, 44, 5, 831-848.
[9] Zouba, M., Baptiste, P., Rebaine, D.(2009)Scheduling identical parallel machines and operators within a period based changing mode. Computers and Operations Research, 36, 3231-3239.