Introduction of Uncertainty into Data Envelopment Analysis

2024 Fiscal Year Final Research Report

• Project/Area Number: 22K14443
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 25010:Social systems engineering-related
• Research Institution: Tokyo University of Science
• Principal Investigator: Zhao Yu 東京理科大学, 経営学部経営学科, 講師 (40879384)
• Project Period (FY): 2022-04-01 – 2025-03-31
• Keywords: 不確実性 / 効率性の評価 / フロンティア推定法 / データ包絡分析法

Outline of Final Research Achievements

This study develops a novel method for estimating production frontiers under uncertainty in multi-input, multi-output systems, based on Data Envelopment Analysis (DEA). Traditional DEA models often fail to adequately account for stochastic fluctuations and measurement errors in the data, resulting in limited accuracy in frontier estimation. To address this issue, the proposed approach integrates statistical learning theory, machine learning, and information theory to enhance the precision and flexibility of both frontier and efficiency estimation. Simulation studies and empirical applications demonstrate that the proposed method outperforms existing DEA models in terms of accuracy and robustness.

Free Research Field

オペレーションズ・リサーチ、統計学

Academic Significance and Societal Importance of the Research Achievements

本研究は、DEAに不確実性を導入することで、従来の決定論的モデルでは捉えきれなかったデータの確率的変動や観測誤差に対応可能な理論的枠組みを提示した点で学術的意義がある。統計的学習理論や機械学習との融合により、効率性評価の精度と汎用性を高めた。また、提案手法は、医療、金融、公共部門など、実社会での意思決定支援にも応用可能であり、限られた資源の有効活用やサービスの質の向上に貢献する社会的意義を有する。