A deterministic approach for performance assessment and optimization of power distribution units in Iran

Abstract

This paper presents a deterministic approach for performance assessment and optimization of power distribution units in Iran. The deterministic approach is composed of data envelopment analysis (DEA), principal component analysis (PCA) and correlation techniques. Seventeen electricity distribution units have been considered for the purpose of this study. Previous studies have generally used input–output DEA models for benchmarking and evaluation of electricity distribution units. However, this study considers an integrated deterministic DEA–PCA approach since the DEA model should be verified and validated by a robust multivariate methodology such as PCA. Moreover, the DEA models are verified and validated by PCA, Spearman and Kendall's Tau correlation techniques, while previous studies do not have the verification and validation features. Also, both input- and output-oriented DEA models are used for sensitivity analysis of the input and output variables. Finally, this is the first study to present an integrated deterministic approach for assessment and optimization of power distributions in Iran.

Introduction

Performance assessment of electricity distribution units is an important issue for decision makers and researchers. There are various methods for estimating efficiency scores of electricity distribution units. These methods are generally classified as deterministic and stochastic methods. Also, one can classify the methods as parametric and non-parametric methods. In the parametric methods, a cost or production function is estimated, whereas in the non-parametric methods, it is not necessary to estimate the cost or production function. Corrected ordinary least squares (COLS) and stochastic frontier analysis (SFA) are parametric models and data envelopment analysis (DEA) and principal component analysis (PCA) are non-parametric models. In addition, COLS, DEA and PCA are the deterministic methods and in contrary, SFA is stochastic method. Most of the efficiency studies of units have been focusing on units within a single country (Försund and Kittelsen, 1998), but a few studies have also compared units from different countries (Jamasb and Pollitt, 2001). There are many studies for estimating efficiency of electricity distributions units which use DEA approach (Jamasb and Pollitt, 2001). DEA is a non-parametric method that uses linear programming to calculate the efficiency in a given set of decision-making units (DMUs). These DMUs utilize a variety of sources as inputs to produce several outputs. Goto and Tsutsui (1998) using DEA model, measured overall cost efficiency and technical efficiency between Japanese and US electricity utilities. Their results showed that Japanese utilities were more efficient than US utilities in terms of technical, allocative and scale efficiency. Försund and Kittelsen (1998) applied DEA efficiency scores to measure Malmquist productivity index in the Norwegian electricity distribution companies. Resende (2002) used non-parametric input-oriented DEA model for evaluation of Brazilian electricity distribution firms. In the Resende study, potential and difficulties with the implementation of yardstick schemes were discussed. Edvardsen and F¢rsund (2003) studied performance of 122 electricity distributors in Denmark, Finland, Norway, Sweden and the Netherlands for the year 1997. They applied input-oriented DEA model and Malmquist productivity index and they found Finnish electricity distributors had most productivity among other countries. Giannakis et al. (2005) applied DEA model to study service quality of UK electricity distribution utilities. They found cost-efficient firms do not necessarily exhibit high service quality and efficiency scores of cost-only models do not show high correlation with those of quality-based models. Estache et al. (2004) applied DEA and econometric methods for performance assessment and ranking of South American electricity companies. They found high correlation between different econometrics’ models and between different DEA models, respectively. However, there was low correlation between DEA and econometrics’ models.

PCA is used to reduce the number of variables under study and consequently the ranking and analysis of DMUs, such as industries, universities, hospitals, cities, etc. (Sharma, 1996; Zhu, 1998; Azadeh and Ebrahimipour, 2002, Azadeh and Ebrahimipour, 2004; Azadeh and Jalal, 2001). Although there is not any study for evaluation of electricity distribution units by PCA, the later method is widely used in other industries (Turker, 1997; Choi and Chu, 2000; Al-Subhi and Al-Harbi, 2000; Rossi and Thomas, 2001; Terziovski, 2002). Zhu (1998) and Premachandra (2001) used PCA as an alternative to DEA. They showed high correlation between DEA and PCA Models. Adler and Golany (2001) have used PCA as a data reduction technique to select inputs and outputs. Cinca and Molinero (2004) have applied PCA to select DEA model.

In this paper both DEA and PCA which are well known deterministic approaches are applied for performance assessment and optimization of electricity distribution units in Iran. Section 2 presents the deterministic approach. Section 3 describes how inputs and outputs are selected. In Section 4 we present the data and corresponding results. Finally, we summarize and conclude the paper in Section 5.