An integrated framework for continuous assessment and improvement of manufacturing systems
Introduction
Major factors influencing the overall productivity of an industrial organization are identified as technology, machinery, management, personnel and rules and procedures [1], [2], [3]. Organizational and management factors plays an important role in the overall performance of manufacturing systems. In fact, managerial and organizational productivity is correlated with the overall performance of a manufacturing system. Furthermore, the overall performance of an industrial organization is often assessed by managerial and organizational productivity. The need for an integrated approach for continuous assessment and improvement of manufacturing systems based on management performance has become essential. Continuous assessment requires manufacturing classifications and taxonomy to be introduced to enhance knowledge and understanding about the behavior of manufacturing systems [4], [5], [6], [7], [8]. Consequently, it will enable predictions to be made about organizational system behavior. In selecting a performance measure or indicator, it is important to consider the measure’s suitability to the control system’s objectives, the measure invasiveness and its complexity [9]. In selecting an appropriate range of performance measures it will be necessary to balance them to make sure various dimensions of manufacturing performance is considered [10], [11]. Furthermore, we need to make sure that one or more dimensions of performance are not stressed to detriment of others.
This study has identified major productivity indicators, which affect management performance in industrial organizations. An integrated study must consider not only the traditional productivity view but also it must consider other views such as efficiency, effectiveness and profitability. Effectiveness is defined as actual output to planned output, efficiency is defined as actual output to actual input and profitability is defined as total revenue to total cost. Furthermore, this study considers the four views of management and organization productivity, which are: (1) traditional productivity, (2) efficiency, (3) effectiveness and (4) profitability. In this study, all of the four views are referred to as management and organization productivity. By consolidating a set of management and organization productivity indicators, the telecommunication sectors may be ranked and analyzed by principal component analysis (PCA). Also, the validity and credibility of the PCA may be verified and validated by numerical taxonomy (NT) approach and non-parametric correlation experiments. It should be mentioned that data envelopment analysis (DEA) was first selected as the verification tool, but several indexes could not be considered due to the unique structure of DEA. Based on examination of 64 plants in Germany, it was concluded that machinery and training play the most important role in productivity improvement of industrial organizations [12]. Hong and colleagues showed that data envelopment analysis (DEA) can be used to evaluate the efficiency of system integration projects and proposed a methodology to overcome the limitations of DEA by utilizing DEA along with machine learning [13]. Multivariate analysis were used with the purpose of identifying critical export marketing success factors by a survey of 134 export activities of manufacturing firms in Denmark [14]. A multivariate analysis was used to test whether there is any relationship between airline flight delays and the financial situation of an airline [15].
Multivariate analysis was used to identify valuation of farmland in Spain [16]. Multivariate analysis was performed with the purpose of identifying critical export marketing success factors [17]. The relative position of United Kingdom car market was assessed with the aid of multivariate statistical analysis [18]. Other researchers used a multivariate linear statistical model to investigate the effects of speed, travel distance and part weight on robot repeatability and accuracy [19]. A fuzzy clustering and classification model for productivity analysis of machinery industry is discussed by Chen and colleagues [20]. A multivariate approach was used among 128 manufacturing organization to indicate that man–machine interfaces are significant contributors to reducing the negative effect of system complexity [21]. Three performance measures, namely, customer satisfaction, productivity and technological competitiveness were collected from a large sample of manufacturing sites in Australia and New Zealand and analyzed by multivariate analysis technique [22]. Application of multivariate techniques including PCA and neural networks in a pulp mill factory is proposed and discussed by Kumar [23]. There are other studies, which show the applications of multivariate analysis in various settings [24], [25].
Section snippets
PCA
Principal component analysis (PCA) is widely used in multivariate statistics such as factor analysis. It is used to reduce the number of variables under study and consequently ranking and analysis of decision-making units (DMUs), such as industries, universities, hospitals, cities, etc. [26], [27], [28], [29], [30], [31], [32], [33]. PCA was applied to selection of monitoring plants for fluoride and two indexes were found [34]. Furthermore, PCA captured the measurement correlations and
Numerical taxonomy
Numerical taxonomy approach is capable of identifying homogeneous from non-homogeneous cases. Furthermore, a group of DMUs by given indexes is divided to homogeneous sub-groups [42]. It also ranks the DMUs in a particular group. The numerical taxonomy approach is as follows [43], [44], [45], [46]:
- Step 1:
Suppose we have k DMUs with p variables (indexes) which can be shown by a k by p matrix X = [xij] where xij is the value of jth index for the ith DMU (i = 1⋯k and j = 1⋯p).
- Step 2:
The k by p matrix is standardized
Integrated framework
To achieve the objectives of this study, a comprehensive study was conducted to locate all economic and technical indicators (indexes), which influence management and organizational performance. These indicators are related to management productivity, efficiency, effectiveness and profitability. Managerial and organizational performances are categorized into four groups: financial, customers’ satisfaction, internal process (including process innovation and production process) and organizational
Verification and validation
To verify the results of PCA, a numerical taxonomy approach must be employed. First, the summary of numerical taxonomy must be obtained. Second, the ranking of the two approaches should be analyzed by Spearman and Kendall Tau correlation experiments. However, it should be identified if cluster analysis is required for numerical taxonomy. Moreover, the distance matrix in numerical taxonomy should be computed and analyzed. As noted di represent the smallest value in each row of the distance
Conclusion
In summary, a unique integrated framework is presented to assess managerial and organizational factors in manufacturing systems. Managers may use this type of modeling approach to assess the performance of various production sites with respect to the management and organizational indicators. In turn, the selected sites would be ranked based on an integrated scientific approach, which reveals the standing of each site with respect to a series of standard management indicators. This would enable
References (73)
Data envelopment analysis vs. principal component analysis: an illustrative study of economic performance of Chinese cities
European Journal of Operation Research
(1998)- et al.
Using principal component analysis in process performance for multivariate data
Omega
(2000) A monothetic clustering method
Pattern Recognition Letters
(1998)Creating highly reliable manufacturing systems: an integrated approach
International Journal of Reliability Quality and Safety Engineering
(2000)- M.A. Azadeh, Creating high reliable manufacturing systems: an integrated approach, in: Proceedings of the Fourth...
- et al.
Facilities Planning
(1996) - et al.
The use of numerical analysis to classify companies according to production control complexity
International Journal of Production Research
(1986) - et al.
A new taxonomy of modern manufacturing practices
International Journal of Operations and Production Management
(1997) An empirical taxonomy of advanced manufacturing technology
International Journal of Operations and Production Management
(2000)Manufacturing classification: lessons from organizational systematic and biological taxonomy
Integrated Manufacturing Systems
(1995)