當應用LED於平面顯示器如燈條(Light Bar)之顯示時，由於LED在色度、亮度與電壓等三大特性上，產出呈常態分佈，一般依其特性區間以組合碼(BIN 碼)代表之。一燈條可從多組LED項目取一所組成，一LED項目可由不同BIN碼所組成，相同LED項目可共用不同BIN碼。因此，在庫存管制上，共用性高之LED BIN料易於使用，而共用程度低之邊BIN料庫存難以去化的問題。尤其LED的跌價損失快，對於多廠區生產時，如何達到跨廠區庫存管制的最佳化實為挑戰。
本研究基於多階層整數規劃法，開發一web化之多階層最佳化LED配置系統(Optimal LED Allocation System, OLAS)，提供生產工單所需之領料配量，以達到LED全產全銷的目標。在計算效益上，本研究提出一基於現有庫存的簡化BIN控制矩陣法，可大幅減少最佳化運算所需之計算時間。在庫存管制上，依低共用性先用的原則，以多階層整數規劃求解，階層零採用從邊BIN碼先用，階層一僅限兩種BIN交錯使用之混打，階層二採混打加僅用單一BIN碼之單打。最後，階層三為補量以預防不良率損失等四大階層，以逐步去化邊餘BIN。此外，並提供多邊顯示的連板倍率限制，符合大尺寸對稱BIN顯示的需求。
結果顯示，就某面板大廠近一個月的使用情況而言，在計算效益上，應用所提之簡化控制矩陣法，可將所需處理之資料量縮減為原先之29%。在求解速度上，應用ILOG引擎求解，平均可於2分鐘內求解完成。在LED庫存管制上，應用四階層的最佳化配量，本系統可持續優先使用低共用性邊BIN碼量。由於LED的全產全銷，更可進一步降低指定LED BIN碼的採購成本。

When adopting LED as lighting source of light-bar for panel display, due to LED characteristics being normal distributed in chromaticity, illumination and voltage, a LED is characterized by BIN codes. In light-bar combination, a light-bar can be combined from one of several LED items; a LED item can be consisted from various BIN codes; the same LED item can use different BIN codes. Hence, high exchangeable BINs are easily used than low exchangeable BIN codes, a.k.a. side-BIN codes, such that a challenge is how to optimize cross-site inventory of side-BINs under fast inventory falling price loss.
This work based on multi-level integer programming method proposes an optimal LED allocation system (OLAS) to allocate LED quantities by work-order, while reaching the objectives of maximizing order fulfillment and minimizing side BINs. In computing efficiency, this work based available stock proposes a simplified bin Control Matrix method to significantly reduce required computation time for allocation optimization. In controlling inventory, according to low exchangeability first policy, the allocating policy by levels of OLAS are as follows: level zero is to use the side BIN codes; level one is to interlace two BIN codes for allocation; level two can allocate single or two interlaced BIN codes to order; and level three reserves extra quantities to prevent production loss. Additionally, this system can satisfy pallet constraints to meet requirements of symmetric BIN codes usages.
After applying OLAS to a panel display company for one month, the result shows that the required data can be reduced more than 71% than the original data amount by using the simplified Control Matrix method; the mean solving time is less than two minutes by utilizing ILOG engine; finally, first allocated the low exchangeable BIN code’s inventory after adopting the four-level integer programming model.

[1] Display research, www.displaysearch.com, 2012.
[2] Linear programming. Available：http://zh.wikipedia.org .
[3] Vašek Chvátal, Linear programming, W. H. Freeman, 1983.
[4] A. Schrijver, “Theory of Linear and Integer Programming”, John Wiley & Sons Ltd, Chichester, England, 1998.
[5] W. F. Bialas and M. H. Karwan, “Two-Level Linear programming”, Management Science, Vol. 30, No. 8, pp. 1004-1020, 1984.
[6] W. F. Bialas and M. H. Karwan, “Multilevel Linear Programming,” Industrial Engineering SUNY at Buffalo, Research Report, no. 78-1, 1978.
[7] Paul Shaw & Constraint Programming Group, IBM ILOG CP Optimizer CP-AI-OR 2009 Master class, 2009.
[8] H.-C. Yang, Y.-L. Chen, M.-H. Hung, and F.-T. Cheng, “Virtual Production Control System,” Proc. of the IEEE Conference on Automation Science and Engineering, pp. 984-989, 2010.
[9] E. A. Demirtas and O. Ustun, “An integrated multi-objective decision making process for supplier selection and order allocation,” Omega, vol. 36, no. 1, pp. 76-90, 2008.
[10] C. Kasemseta and V. Kachitvichyanukul, “Bi-level multi-objective mathematical model for job-shop scheduling: the application of Theory of Constraints,” International Journal of Production Research, Vol. 48, Iss. 20, pp. 6137-6154, 2010.
[11] R. J. Kuo and Y. S. Han, “A hybrid of genetic algorithm and particle swarm optimization for solving bi-level linear programming problem– A case study on supply chain model,” Applied Mathematical Modelling, vol.5, Iss. 8, pp. 3905-3917, 2011.
[12] Y.-H. Lin, J.-R. Shie, and C.-H. Tsai, “Using an artificial neural network prediction model to optimize work-in-process inventory level for wafer fabrication,” Expert Systems with Applications, vol. 36, pp. 3421-3427, 2009.