[[alternative]]A Study on Determining Recycling and Treatment Fees to Recycling Industries for Waste Computer Appliances-- an Applications of Multi-Level Stochastic Programming

計畫編號：NSC99-2221-E032-025-MY2研究期間：20110801~20120731研究經費：599,000[[abstract]]本計畫旨在利用多階層隨機規劃 (multi-level stochastic programming) 模型，求解環境保護議題中之資訊廢棄物回收處理補貼費率制定的問題。資訊廢棄物為當今環境保護工作中最為關注的對象之一，環境保護署基金管理委員會向資訊責任業者(製造商與進口商) 收取處理費用，用以補貼回收處理業者以期提高回收率，俾利污染控制以及資源再利用。其中補貼費率的制定涉及到多個利益關係人的權益以及整體資源之合理分配，而多階層規劃將為一描述其間互動關係的可行模型。另一方面，資訊物品雖有其使用年限 (譬如桌上型電腦估計使用五年)，但廢棄物的淘汰時機會受到個人習慣、經濟與科技因素影響，故與回收率具不確定性，因而將為一多階層隨機規劃模型。再者，資訊廢棄物回收後，除了進入回收體系外，部份具有剩餘價值者將會進入二手市場，此二手市場的銷售狀況也會受到經濟與科技因素的影響，而形成另一不確定性因子。由於此模型中的隨機性及非線性並存，造成求解的困難，本研究將嘗試使用傳統方法與基因演算法分別解決。 本計畫預計分為三個主要議題： 一、資訊廢棄物回收處理補貼費率制定的多階層隨機(非線性)規劃建模與求解。 二、考量二手市場影響的兩階層分散式隨機(非線性)規劃建模與求解。 三、建立基因演算法的解算步驟與求解。 以上三個部份為本計畫的重點，期盼藉此整合型計畫的長期發展，解決一些環境保護的管理決策問題；並期於一些整合關鍵點上突破，產生學術上的貢獻。[[abstract]]This study attempts to optimize the resource allocation of recycling fund, controlled by Recycling Fund Management Board (RFMB) in the Environmental Protection Administration of ROC Government, through the decision of subsidy rate of the electronic waste to recycling industry in Taiwan. Electronic waste is the major concern of environmental protection for toxic control and resources recovering. RFMB obtains processing fee from electronic industries and importers and gives subsidy to recycling industries for increasing recycling rate or amount. The interactive relationship among stakeholders can be formulated by multi-level stochastic programming (MLSP), considering the uncertainty of end-of-life of PCs and recycling rate. The recycling rate will be affected by the factors of consumers' behavior, economic and technological factors. In addition, there is a second-hand market that will affect the recycling amount of electronics based on the economic considerations on used PCs and their components. Thus, an extension of the bi-level decentralized stochastic nonlinear model will be established. Due to the uncertainty and non-linearity in the model, we try to use genetic algorithms as an option for solving the problem. Some comparisons between genetic algorithms and traditional approach are also made. Three major issues in the project are listed as follows. 1. Solving the problem of the bi-level stochastic (nonlinear) programming for deciding subsidy and treatment fee of the electronic waste to recycling industries. 2. Establishing a bi-level decentralized model considering the effect of second-hand market and the role of electronic industries and importers. 3. Solving the extended model by the traditional approach and genetic algorithms. Hope the project is supported for the benefit of academic development and problem-solving in environmental protection issues of Taiwan.[[sponsorship]]科技部[[notice]]補正完