本研究以「結合 AIoT 與 ABC 作業成本制之智慧物流投資效益與決策架構：以都市倉儲系統為例」為題，旨在建構一套整合人工智慧物聯網（AIoT）與作業成本制（ABC）的智慧決策模型，以改善企業成本資訊延遲與投資報酬評估不確定性之問題。研究採理論建構與試算驗證並行設計，並以都市倉儲產業為場域，試算比較傳統作業流程與導入 AIoT × ABC 系統後之成效。試算結果顯示，AIoT × ABC 模型能顯著改善成本結構與作業效率。單位間接成本由 6.04 元降至 4.89 元，降低約 19%；平均挑貨時間由 1.5 分鐘縮短至 0.9 分鐘，效率提升約 40%；作業錯誤率由 4.2% 降至 1.3%，減少約 69%；貨架利用率由 68% 提升至 84%。雖能源消耗略增 19%，但整體能源使用效率仍提升。ROI 試算呈穩定成長，於第四週與對照組相比獲利差距平均達 60 元，顯示智慧化投資具明顯回報潛力。灰色關聯分析顯示，能源效率、設備稼動率與作業週轉時間為影響 ROI 的主要因素，其關聯係數均高於 0.79。研究同時提及「動態成本池」與「智慧決策支援理論（Smart Decision Support Theory, SDST）」，以拓展 ABC 模型於即時資料環境之應用。研究結果顯示，AIoT × ABC 模型可作為企業智慧決策與永續投資的重要理論與實務依據。;This study, titled “A Study on the Investment Efficiency and Decision-Making Framework of Smart Logistics Based on AIoT and Activity-Based Costing: A Case of Urban Warehouse Systems” aims to establish an intelligent decision-making model that integrates the Artificial Intelligence of Things (AIoT) with Activity-Based Costing (ABC). The model seeks to address delays in cost information and uncertainties in investment evaluation through data-driven decision support. A mixed-method design combining theoretical modeling and simulation was adopted. Using the urban warehousing sector as the test field, the study compared traditional operations with the AIoT × ABC integrated system. Simulation results demonstrated significant improvements: unit indirect cost decreased from 6.04 TWD to 4.89 TWD (a 19% reduction), average picking time shortened from 1.5 to 0.9 minutes (a 40% improvement), error rate fell from 4.2% to 1.3% (a 69% reduction), and shelf utilization increased from 68% to 84%. Although energy consumption rose by 19%, overall efficiency improved. The ROI showed consistent upward growth, with a profit gap of approximately 60 TWD by the fourth week, indicating a clear potential for return on investment. Grey Relational Analysis (GRA) identified energy efficiency, equipment utilization, and process turnaround time as key determinants of ROI, each with a correlation coefficient above 0.79. The Dynamic Cost Pool and Smart Decision Support Theory (SDST) were proposed to extend the applicability of ABC in real-time data environments. The results confirm that the AIoT × ABC model provides a valuable theoretical and practical foundation for intelligent decision-making and sustainable investment strategies.
