應用序優化理論於組裝式生產系統的利潤最佳化

Apply ordinal optimization theory to optimize the profit of assemble-to-order systems

徐士堡(Shih-pao Shu)

指導教授 ： 洪士程

朝陽科技大學/資訊學院/資訊工程系/碩士(2012年)

摘要

在本篇論文中，提出一個結合支持向量回歸(SVR)、基因演算法(GA)和最佳資源預算分配(OCBA)的演算法，用來解決從巨大的設計變數空間中，找到一個最佳的變數組合，來達到最好目標函數值的整數型變數組合最佳化問題。雖然文獻中已經有許多的全域最佳解搜尋方法來解決此類的問題，但是需要大量的世代搜尋，會耗費許多的計算成本及時間，因此，在本篇分為兩個階段，首先在第一階段使用基因演算法搭配近似模型支持向量回歸，來評估適應值並選出足夠好的解之子集合，接著第二階段使用最佳資源預算分配的方法，透過合理的分配計算量，對於關鍵的設計變數分配較多的計算量，而不重要的設計變數則分配較少的計算量，如此對於真正好的設計變數可獲得更大的正確選擇機率，並且也大大地減少了計算的複雜度。我們將提出的方法應用在製造工業應用上的組裝式生產系統(ATOS)上的利潤最大化問題，是一種整數型變數組合的最佳化問題，利用提出的演算法求取最佳的零件庫存儲備預設值，使得組裝式生產系統可以得到最大利潤。最後將本篇方法得到的結果與其他方法做比較，經過數次模擬結果，本篇方法表現超越其他方法。

關鍵字

最佳資源預算分配 ； 基因演算法 ； 支撐向量回歸 ； 組裝式生產系統

並列摘要

In this thesis, a method combined the support vector regression (SVR), genetic algorithm (GA) and optimal computing budget allocation (OCBA) is proposed to solve the combinatorial optimization problems with huge size of discrete solution space. The goal is to find an optimal solution which achieves the best objective function value within a reasonable computation time. Although there are many heuristic methods, they need more computation time to obtain the optimal solution. To overcome this drawback, the proposed method is divided into two stages. In the first stage, the GA with SVR approximation model is used to assess the fitness value and select a good enough solution subset from entire discrete solution space. In the second stage, we use the OCBA to greatly reduce the computational complexity. Through a reasonable allocation of computation, more computations are allocated for the better design variables such that they can get more correct choice probability. The proposed method is applied to the assembly-to-order (ATO) system which is formulated as a combinatorial optimization problem with integer variables that possesses a huge solution space. The proposed method is tested on an ATO system comprising 10 items on 6 products for determining a good enough target inventory level using limited computation time such that the expected total profit per period is maximized. Finally, the solution quality is demonstrated by comparing with those obtained by other competing methods.

並列關鍵字

assembly-to-order ； Genetic Algorithms ； optimal computing budget allocation ； support Vector regression

參考文獻

[3] S.J. Abspoel, L.F.P. Etman, J. Vervoort, R.A. van Rooij, A.J.G. Schoofs, J.E. Rooda, “Simulation based optimization of stochastic systems with integer design variables by sequential multipoint linear approximation”, Structural and Multidisciplinary Optimization, vol. 22, no. 2, pp. 125-138, 2001.

[4] Y. Xiao, J. Chen, C.Y. Lee, “Optimal decisions for assemble-to-order systems with uncertain assembly capacity,” International Journal of Production Economics, vol. 123, no. 1, pp. 155-165, 2010.

[5] S.S. Ko, J.Y. Choi, D.W. Seo, “Approximations of lead-time distributions in an assemble-to-order system under a base-stock policy,” Computers & Operations Research, vol. 38, no. 2, pp. 582-590, 2011.

[6] G. L. Nemhauser, A.W. Laurence, Integer and Combinatorial Optimization. John Wiley & Sons, New York,1999.

[7] L. Ntaimo, S. Sen, “The million-variable march for stochastic combinatorial optimization,” Journal of Global Optimization, vol. 32, no. 3, pp. 385-400, 2005.