مدل های کمی برای سیستم های سنجش عملکرد — ملاحظات متناوب
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|10703||2003||10 صفحه PDF||سفارش دهید||4434 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : International Journal of Production Economics, Volume 86, Issue 1, 11 October 2003, Pages 81–90
This paper revisits the recent works of Suwignjo, Bititci and Carrie (SBC) (Int. J. Prod. Econom. 64 (2000) 231) and Bititci, Suwignjo and Carrie (BSC) (Int. J. Prod. Econom. 69 (2001) 15). We show how a generalizable analytical hierarchy technique based on Saaty's systems-with-feedback approach, also known as the analytical network process (ANP), can be applied, as an alternative methodology, to SBC's quantitative model for performance measurement system. The proposed enhancement to SBC's approach can be completed through the utilization of a supermatrix that can arrive at a solution where the necessary combined weights of the factors influencing the performance measure are a result. In addition, we also show how the ANP approach could be used to enhance the dynamic evaluation of BSC's manufacturing strategy performance evaluation model.
Suwignjo, Bititci and Carrie (SBC; Suwignjo et al., 2000) and Bititci, Suwignjo and Carrie (BSC; Bititci et al., 2001) provide an innovative framework and supporting system that allows organizations to incorporate and map performance measures in a hierarchical fashion. Central to this approach is the development and application of their tool defined as the quantitative model for performance measurement system (QMPMS) that relies on the analytic hierarchy process (AHP) to quantify factors (tangible and intangible) for performance. They decompose their process into three steps (p. 231 of SBC): 1. identification of factors affecting performance and their relationships, 2. structuring the factors hierarchically, 3. quantifying the effect of the factors on performance. These three steps are appropriate for use within an AHP framework as described by Saaty (1980) and Saaty (1996). Yet, the hierarchy they form in SBC is more of a network hierarchy, which incorporates a number of inter-relationships. Using this network formation, we recommend the use of a technique developed by Saaty (1996) that incorporates various “feedbacks” for the generation of a stable set of weights incorporating the three effects detailed by SBC. This feedback model has also been defined as the analytical network process (ANP) (Hamalainen and Seppalainen, 1986; Saaty, 1996). SBC do mention the use of ANP for reducing the rank-reversal problem. Yet, the advantages of the ANP approach goes even further allowing for a direct calculation of the combined effects of all the factors, utilizing a Markovian process and a more complete set of relationships that are allowed to flow through the network. To show how these inter-relationships can be modeled, we will review the various relationships (effects), their formation into a supermatrix and then their calculation. This alternative approach can provide a single coherent model without the many, separate identification and iterations of various hierarchies, paths, and detailed factor aggregations. In addition, Bititci et al. (2001) (BSC), apply their QMPMS technique for a “dynamic environment” which can be used for manufacturing strategy evaluation and management. We also extend the BSC model for manufacturing performance evaluation to incorporate the feedback mechanisms that can form one coherent long-term strategy for the organization as the current and future dynamics are considered. Thus the contributions of this paper are to: (1) show how the supermatrix approach can be applied to the QMPMS process with fewer requirements of path (or cognitive map) identification through one aggregate model, and (2) show an alternative modeling of the dynamic nature of strategic decisions based on performance measurement.
نتیجه گیری انگلیسی
In this paper, we have shown an alternative method to quantify the combined effects of factors on organizational performance measures using the supermatrix approach. The technique still has some of the difficulties of the AHP approach mentioned in SCB, such as the perceptual subjectivity of decision-maker input and slight problems with rank reversal, even though some have posited that the rank-reversal problem can be mitigated through the supermatrix approach (Saaty, 1994; Schenkerman, 1994). The advantages of the supermatrix approach are at least two-fold. Firstly, it can be easily altered for dynamic environments. The number of factors can be easily added and removed by either adding or removing columns/rows for factors, while relationships (arcs) can be easily removed or added by setting values to zero or non-negative values. Paths and accumulations do not have to be determined for indirect effects. There is also one series of calculations that has to be completed, instead of a series of calculations for each factor or for multiple time periods, where paths have to be determined and data has to be stored. Multiple time periods may also be an issue for a single hierarchy as the QMPMS approach which would need to be aggregated across these time periods. Thus, this simplification in modeling and quantification can prove beneficial to programs such as QMPMS the dynamics of the environment explicitly in a single model instead of a sensitivity analysis that requires changes in factors overtime. A major difficulty of the ANP approach is that additional interdependency relationships increase (geometrically) the number of pairwise comparison matrices and pairwise comparison questions required for an evaluation. For example, in the manufacturing strategy selection problem of BSC, they required only eight pairwise comparison matrices and 25 pairwise comparison questions. In the ANP model presented here, 20 pairwise comparison matrices with 73 pairwise comparison questions are required. But the ANP model does provide integration of the multiperiod planning horizon that was not evident in the original model. Also, QMPMS required calculations for inherent effects, self-interaction effects, horizontal effects, and horizontal effects through other factors calculations. In addition, aggregation over multiple time periods using QMPMS analyses have typically required additional aggregating models such as linear and goal programming. Varying some of the factors and scorings within the supermatrix can complete sensitivity analysis of this problem. Selecting which factors to vary is a managerial decision and we do not introduce this issue here, where the focus is on the implementation of the technique within a QMPMS-like tool.