考慮最佳生產規模與環境效益之隨機產能規劃

Abstract

隨著全球環保意識的提升，對於製造廠商而言，在進行產能規劃決策時考慮非意欲產出（Undesirable Output）的有害影響變得更加重要。同時，確保這些廠房以最佳生產規模運作並滿足市場需求亦做為產能規劃問題中的重點之一。然而，對於製造廠而言，處理生產力、永續性和滿足需求之間的權衡關係則相當困難。此外，市場需求可能會隨著時間出現波動，使得精準地縮小供需差距變得更有挑戰性。本研究利用資料包絡分析法（Data Envelopment Analysis）建立數學模型，試圖提供一個同時考慮生產力、永續性以及滿足市場需求的目標，讓市場中的廠商能夠參考並改善其資源分配。本研究進一步應用了穩健最佳化方法（Robust Optimization）來處理市場需求的不確定性。最後，本研究透過資料模擬驗證所提出框架之有效性，並呈現該框架在資料中存在大幅度地的變異時，相較於未使用穩健最佳化方法的框架可降低約 22% 的完全資訊期望值（Expected Value of Perfect Information），顯示該框架擁有良好處理不確定性的能力。本研究結果可做為企業在面臨產能規劃問題時的參考依據，使其能夠建立一個全面的目標並根據該目標改善資源分配。

With the increasing global environmental awareness, considering the harmful effects of byproducts when making capacity planning decisions has become more crucial for manufacturing plants. At the same time, operating with optimal efficiency and meet the market demands is of significant importance for these plants. However, the trade-off relationship between productivity, sustainability, and demand fulfillment can be complicated for plants to manage simultaneously. Moreover, market demands may fluctuate over time, making it even more challenging to reduce the gap between supply and demands. This study thus proposes a mathematical model that incorporates data envelopment analysis constraints to formulate an efficient target representing the trade-off between the most productive scale size, most sustainable scale size, and demand fulfillment. This study also applies a robust optimization approach to address demand uncertainty. The results demonstrate how manufacturing plants should allocate resources based on a given set of inputs, outputs, and undesirable outputs. The computational study chapter validates the effectiveness of the proposed framework and shows that it can reduce the expected value of perfect information by 22% comparing to a naïve approach framework. The findings provide detailed guidance on how firms can operate under capacity planning problems from a comprehensive perspective.