Performance Measurement for Network Structured Operation Processes
LIU Fuh-Hwa Franklin
|摘要:||近來績效評估相關文獻中紛見的問題為評量單位的作業流程以縱列或網路型式，他們應用許多不同的解題技術來解決此類問題。這些與實務更接近的不同問題敘述，與經研發的不同理論解題方法，兩者所組合而成的新發展，達到理論與實務之並進。在多指標績效評量中，以無母數分析方法制定一組評量指標的權重。其中資料包絡分析法有放射型、多向性及差額變數模式(SBM)，評量出一群互比的決策單位其中每個決策單位的最佳績效。另外，在組織決策者管理其下許多作業單位時，評量此決策者的績效表現可應用共同權重分析(Common Weights Analysis, CWA)或最妥協權重分析(The Most Compromising Set of Indices ofWeights Analysis, MCWA)求得。本計畫分A、B、C三子題，各子題由三小題構成。子題A：以差額變數之資料包絡模型求解縱列式流程中各單位之相對績效，小題A1為兩階段流程； A2為多階段流程； A3為兩階段共用資源流程。子題B：以多向性之資料包絡模型評量各單位之相對績效，小題B1為量測超高效程度； B2為多向性模型之延伸；B3為以B2評量兩階段流程。子題C：具網路式流程之群體其績效指標之共同權重解析，小題C1為縱列式兩階段流程；C2為並列式兩階段流程；C3為網路式流程。子題A、B及C無論是問題本身類型或是解決方法都彼此相關。整體計畫針對網路式結構問題、資源分享以及產出限制作探討，以差額模式、MEA、CWA及MCWA來分析。|
Performance measurement problems with operation processes in tandem or in network structures are appeared in recent literature. Varieties problem solving approaches also applied to those problems. The combinations of problems and solving approaches lead the theoretical achievement to real application. Given the multiple criteria performance measurement, non-parametric analysis models provide the set of weights of those criteria. Data envelopment analysis models could be radial-based, multi-directional efficiency analysis, and slack-based to measure the efficiency for each decision-making unit against its peer units. On the other hand, in the case that a general decision-making of an organization governs several operation units, one is to measure the best-practice efficiency of the general decision-maker. A common set of weights of the criteria could be calculated by employing common weight analysis (CWA) or most compromised weight analysis (MCWA). This proposal contains three projects A, B and C. Each project has three topics. Project A is entitled as: Assessing performance of operation units in tandem processes with slack-based DEA models. The three topics are: (A1) two processes, (A2) multiple processes, and (A3) two processes with shared resources and outputs. Project B is entitled as: Performance assessment with Multi-directional Efficiency Analysis (MEA) DEA models. The three topics are: (B1) Measure the super efficiency, (B2) extended MEA model, and (B3) employ (B2) for two stages processes. Project C is entitled as: Determine the common weights of performance indices for a group of units in network structured processes. The three topics are: (C1) two processes in tandem, (C2) two processes in parallel, and (C3) network structured processes. A, B and C projects are related either by their problem types or by their problem solving approached. The entire proposal is aimed to the network structured problems, and further with shared resources and outputs side conditions. Those problems would be approached by slack-based, MEA, CWA, and MCWA mathematical models.