以多準則決策進行可程式邏輯控制器之評選

Abstract

本研究根據專家經驗並採用層級分析程序法（Analytic Hierarchy Process, AHP）及決策實驗室分析法（Decision Making Trial and Evaluation Laboratory, DEMATEL）歸納計算出可程式邏輯控制器（Programmable Logic Controller, PLC）各指標因素的因果關係及權重排序，並以前三大市佔率的可程式邏輯控制器方案（Allen-Bradley、Mitsubishi 及 SIEMENS）作評比並提出建議，以作為日後監控系統評選可程式邏輯控制器的決策依據。此外，本文另以雷達圖呈現，視覺化三個方案，即 Allen-Bradley、Mitsubishi 及 SIEMENS 的強項和弱項供使用者參考。此次研究採用層級分析程序法結合決策實驗室分析法，不僅減少研究過程中複雜的層級比較，同時倚重專家知識對於屬性間的影響因素判定，使研究結果更為精確。最初，層次分析法被應用到終端使用者導出評估屬性的獨立權重。之後，決策實驗室分析法被應用到代理商或系統商計算求的評估屬性間的因果關係，以取得相依權重。相較過去對可程式邏輯控制器的評選大多採用少數專家主觀偏好，本研究則是透過較系統性的科學方式來解決評選問題。方案評比結果指出功能面及品質面，Allen-Bradley 評價勝於其他兩家；Mitsubishi 則是在成本面上滿意度最高，遠勝於其他兩家；服務面及運轉面上，SIEMENS 滿意度高於其他兩家。

In this thesis, AHP (Analytic Hierarchy Process) and DEMATEL (Decision Making Trial and Evaluation Laboratory) are fused to evaluate Programmable Logic Controllers (PLC). Initially, AHP is applied to end-users for deriving the independent weights of evaluation criteria (associated with sub-criteria). Thereafter, DEMATEL is applied to franchisees for calculating the causalities among the criteria to obtain the interdependent weights. Furthermore, a radar plot is displayed to visualize the strengths and weaknesses of three alternatives, namely, Allen-Bradley, Mitsubishi and SIEMENS. Rather than depending on experts’ subjective preferences, obviously, this thesis provides a systematic framework to carry out PLC assessment in practice. More importantly, with incurring limited pair-wise comparisons, the presented framework is highly feasible to perform PLC recommendation for fulfilling factory monitoring.