Available online at http://www.urpjournals.com International Journal of Fiber and Textile Research Universal Research Publications. All rights reserved ISSN 2277-7156 Original Article AN APPROACH FOR ACCEPTING/REJECTING CONSISTENCY OF ANALYTIC HIERARCHY PROCESS AS A CRITERION DETERMINING TECHNOLOGICAL VALUE OF EGYPTIAN COTTON Khaled M. Hussein*1, Ibrahim A. M. Ebaido1, Rokaya.M.Hassan1 and Hayam S. A. Fateh2 1 Cotton Research Institute (CRI), Agricultural Research Center (ARC), Giza, Egypt Cent. Lab. For Design and Stat. Anal. Res., Agricultural Research Center (ARC), Giza, Egypt Email: [email protected]* 2 Received 10 January 2013; accepted 11 February 2013 Abstract Two statistical approaches for accepting/rejecting consistency of analytic hierarchy process (AHP) were used in this paper. First approach was the traditional and follows Consistency Ratio (CR); second approach was the new and follows Random Index (RI). By the two approaches we had acceptance consistency in pair-wise reciprocal comparison in the analytic hierarchy process (AHP) as a quality criterion determining the technological value of cotton. The highest technological value belongs cotton variety Giza 87 (16.97) and the lowest value belongs cotton variety Giza 90 (12.87). That is accepted due to the increase of weights of the main determinant for technological value criterias, i.e. fiber length (UHML) and strength (FS). The high values of spearman and partial correlations of fiber length and strength with yarn strength verify the accuracy of these variables in modeling AHP. Values of partial correlation of micronaire value with yarn strength reveal the effect of yarn count of yarn strength. The effect of short fiber content on yarn strength seemed to be inter-correlated with other variables. Generally, partial correlation values verified the accuracy of weights fiber properties modeling AHP, except for fiber elongation which showed insignificant either correlation or partial correlation with yarn strength. Therefore, it is considerable to release fiber elongation of AHP equation in other run. © 2013 Universal Research Publications. All rights reserved Keywords: Consistency, Analytic Hierarchy Process, Random Index, Technological Value. 1. INTRODUCTION Definition and fulfillment of the Analytic Hierarchy Process (AHP) made by [1] and [2], as a Multi Criteria Decision Making (MCDM) technique that represents a complex decision problem as a hierarchy with different levels. Each level of the Analytic Hierarchy Process (AHP) contains different elements with a relevant common characteristic, hence, to determine the degree of quality requirements achieved in the Analytic Hierarchy Process (AHP), [3], recommend using AHP a cardinal measure of the importance or priority of each element in a level by pair-wise comparisons of all elements in that level. Each element in every level of the Analytic Hierarchy Process (AHP) serves as the basis for effecting pair-wise comparisons of the elements in the immediate lower level of the hierarchy as noted both [4], [5] and [6]. They also added the final priorities of the elements in the lowest level (decision alternatives) obtained using the principle of 1 hierarchical composition, and consequently these lead to the overall ranking of design alternatives. The pair-wise reciprocal comparison matrices in the analytic hierarchy process (AHP) has been used by [7] and [8], in a trilogy reciprocal comparison matrix to determine the technological value of upland cotton. Whilst, [9] and [10], used the pair-wise in a trilogy reciprocal comparison matrix and in a pentagonal matrix to determine each of technological and marketing value of Egyptian cotton. The approach for accepting/rejecting consistency of the pair-wise reciprocal comparison matrices in the analytic hierarchy process (AHP) by [1], [11], [4], and [2], was depends on result of the consistency ratio (CR). The same approach has been used by [7], [8], [9], and [10]. [12], have studied the consistency in random pair-wise reciprocal comparison matrices which is the heart of the Analytic Hierarchy Process (AHP) of different sizes and they reached to different statistical approach for International Journal of Fiber and Textile Research 2013; 3(1): 1-5 accepting/rejecting consistency of the pair-wise reciprocal comparison matrices depends on Random Index (RI). Therefore, this paper presents a comparison between two approaches for accepting/rejecting consistency of the pairwise reciprocal comparison matrices used in Analytic Hierarchy Process (AHP) as a quality criterion determining technological value of the Egyptian cotton, as well as the accuracy of weights criteria resulting from consistency. 2. MATERIAL AND METHODS The material used in this study included the following Egyptian cotton varieties, Giza70, Giza87, Giza88, and Giza92 representatives of extra long Egyptian cotton category, while the long staple Egypt class was represents at Giza80, Giza86, and Giza90. Of each variety of Egyptian cotton have been taking 3 levels of quality expressed in Egyptian grading system in the following grades, G/FG ( Good to Fully Good), G (Good) and FGF/G ( Fully Good Fair/ Good). The samples of those varieties and their grades been taken from season 2011. Fiber upper half mean length (UHML), uniformity index (UI), micronaire value (MIC), fiber strength (FS) and fiber elongation (FE %) were all determined on the High Volume Instrument (HVI) according to [13]. Further the Sutter Web Comb Sorter been used to determine short fiber content by weight (SFC %) as directed in the [13]. The lint cotton samples were spun into the two-carded ring counts number (Ne) 40 and 50 using the 3.6 twist multiplier. Carded ring yarn skein strength (lea product) was measured according to [13]. The measurements of the materials characterization used in the present study were under controlled atmospheric conditions (65 to 75 F° temperatures and 63 to 67% relative humidity), due to conducted at the laboratories of the Cotton Research Institute, Agricultural Research Center, Giza - Egypt. Collected data were subjected to the proper of statistical analysis of spearman and spearman partial correlations according to the procedure described by [14]. The data were statistically analyzed by using the computer statistical software package SAS statistical software [15]. Accepting/rejecting consistency 2.1. saatys approach [8], and [9], introduced the pair-wise reciprocal comparison matrices in trilogy matrix, and both of the pair-wise was as a quality criterion determining the technological value of cotton by using the analytic hierarchy process (AHP). However [8], determined the technological value of the upland cotton, while [9], determined the technological value of the Egyptian cotton, so the weights of the criteria used in the determination process was different between both [8], and [9]. Table (1), presents review of the pair-wise reciprocal comparisons in the form of trilogy matrix specific each of [8], [9], moreover, Consistency Ratio (CR) and the criteria weights obtained by both of them. Accepting/rejecting consistency in the analytic hierarchy process (AHP) according to saatys approach was depending on apportionment Consistency Index (CI) and Random Consistency Index (RCI) ≤ 0.1 consistency ratio (CR). 2 If the value of CR is 0.1 or less, then the pair-wise reciprocal comparison matrix is considered to be consistent and acceptable, otherwise the decision maker has to make some changes in the entry of the pair-wise comparison matrix, [1]. From results of consistency ratio (CR), we can sum up that, saatys approach showed a consistent pair-wise comparison by a 0.006 (CR) for Hussein and 0.003 (CR) for Majumdar, these values of CR is less than 0.1, consequently each of the MIAHP equation whether pertains Hussein or Majumdar fits as a quality criterion determining the technological value of cotton. 2.2. Alonso and Lamata approach [12], derived the following approach and applied it through a large number of generated matrices with different sizes. They exploited the results to show the acceptance maximum positive eigen-value (λ max) for various sizes of matrices and under different levels of probability (α). Finally, they found goodness of fit that regressed the λ max values on their corresponding matrix order (matrix size) (n). Where the linear relation between λ max values and the matrix order (n) was more valid and accurate with a correlation coefficient being r = 0.99**, consequently the final product was depending on a consistency index (λ max) and level of consistency needed (α) 0 < a ≤ 1. In fact, this method is very simple criterion for accepting/rejecting matrix; furthermore, this method is able to test large sizes of matrices under different levels of probability. In addition, this level provide adaptability to different scopes, accordingly, we can decide if a specific matrix is a sufficiently consistent matrix (or not) due to the [12], approach in accepting/rejecting consistency by using their equivalent: RandomIndex(RI) = λMax N + α(1.7699N - 4.3513) , where ( α ) is the probability value, N is the matrix order (size of the matrix), 1.7669 is the least- square and -4.3513 is the regression constant. Table 2, clarifies that [12], approach consider to be a new criterion for acceptance and a new index for representing consistency in pair-wise reciprocal comparison matrices, hence, this index and criterion allows the decision maker to study the consistency of each matrix in an adaptable way. The decision maker when using the Random Index (RI) (maximum positive eigen-value - λ max) in different levels of α can decide about the matrix consistency using not only the matrix entries but also the level of consistency that he needs in this particular case. Generally, due to accepting / rejecting the pair- wise reciprocal comparison matrices consistency in the analytic International Journal of Fiber and Textile Research 2013; 3(1): 1-5 Table (1), Review each of pair-wise comparison matrix, Consistency Ratio (CR), as well as criteria weights obtained by Majumdar and Hussein Pair-wise comparison matrix of criteria according to Majumdar Criteria Tensile Length Fineness Tensile 1 1/2 3 Length 2 1 5 Fineness 1/3 1/5 1 Consistency Ratio (CR) = 0.003 (Accepted) criteria weights obtained by Majumdar Length UHML UI SFC 0.291 0.145 0.145 Tensile FS FE 0.270 0.039 Fineness FF 0.11 Pair-wise comparison matrix of criteria according to Hussein Criteria Tensile Length Fineness Tensile 1 1 7 Length 1 1 9 Fineness 1/7 1/9 1 Consistency Ratio (CR) = 0.006 (Accepted) criteria weights obtained by Hussein Length UHML UI SFC 0.380 0.054 0.054 Tensile FS FE 0.394 0.056 Fineness FF 0.059 Table (2). Accepting/rejecting consistency according to random index (maximum positive λ max), under different levels of α Random Index (RI), according to Alonso and Lamata Maximum positive eigen-value (λ max), Majumdar Maximum positive eigen-value (λ max), Hussein 3.004 3.007 Random Index (RI) at α (0.10) = 3.095 Random Index (RI) at α (0.10) = 3.095 (Accepted) (Accepted) Random Index (RI) at α (0.08) = 3.076 Random Index (RI) at α (0.08) = 3.076 (Accepted) (Accepted) Random Index (RI) at α (0.05) = 3.047 Random Index (RI) at α (0.05) = 3.047 (Accepted) (Accepted) Random Index (RI) at α (0.01) = 3.009 Random Index (RI) at α (0.01) = 3.009 (Accepted) (Accepted) RandomIndex(RI) = λMax N + α(1.7699N - 4.3513) Table (3). Description of the technological characteristics Egyptian cotton varieties, as well as their Mi AHP quantitative values Variety The technological characteristics The technological value UHML UI SFC FS FE FF Mi AHP, H Min 33.8 80.1 9.8 42.4 7.0 3.5 15.56 Giza 70 Mean 34.1 82.8 12.4 44.0 7.2 3.8 15.57 Max 34.5 85.0 14.5 47.1 7.5 4.2 15.82 Min 30.0 78.2 13.7 34.2 7.3 3.5 13.37 Giza 80 Mean 30.7 82.8 15.9 36.6 7.8 4.0 13.63 Max 31.5 86.0 18.2 38.5 8.2 4.6 13.81 Min 32.0 79.2 12.0 39.8 7.0 3.6 14.67 Giza 86 Mean 32.9 83.5 13.4 40.9 7.3 3.9 14.83 Max 33.8 87.2 14.9 42.4 7.6 4.3 15.03 Min 33.7 82.1 6.0 46.7 7.2 3.4 16.60 Giza 87 Mean 34.8 84.7 8.3 48.6 7.4 3.6 16.72 Max 36.5 88.1 11.0 51.0 7.6 4.1 16.97 Min 34.0 80.8 9.0 46.2 7.0 3.5 16.21 Giza 88 Mean 35.4 84.8 10.4 47.4 7.1 3.9 16.43 Max 36.2 87.0 11.8 48.9 7.5 4.2 16.56 Min 28.8 75.2 16.2 32.6 7.0 3.2 12.87 Giza 90 Mean 29.5 80.3 17.8 34.6 7.6 3.7 13.10 Max 30.8 85.0 20.5 36.6 8.3 4.5 13.34 Min 34.0 82.0 5.5 47.1 7.0 3.5 16.54 Giza 92 Mean 34.7 85.2 8.2 48.1 7.3 3.6 16.67 Max 35.1 89.2 11.3 49.3 7.6 3.9 16.79 Mi AHP, H = Multiplicative Analytic Hierarchy Process, [9]. UHML = Upper Half Mean Length, UI = Uniformity Index, SFC = Short Fiber Content, FS = Fiber Strength, FE = Fiber Elongation and FF is the Fiber Fineness expressed by Micronaire reading (MIC). reciprocal comparison matrices consistency in the analytic hierarchy process (AHP), by using the two approaches both of [1], and [12]. We conclude that, acceptance consistency in pair-wise reciprocal comparison by both [8] and [9], in the analytic hierarchy process (AHP). Hence, this acceptance qualifies the MiAHP (Multiplicative Analytic Hierarchy Process) a quantification method the quality of cotton in the form of technological value. 3 3. RESULTS AND DISCUSSION 3.1. Quantification the technological value of Egyptian cotton Results tabulated in Table 3 indicates that, the highest technological value belongs Giza87 (16.97) and the lowest technological value belongs Giza90 (12.87). That has accepted due to increase the weights of criterias length (UHML) and strength (FS) were 0.380 and 0.394, International Journal of Fiber and Textile Research 2013; 3(1): 1-5 Table 4 . Spearman and spearman partial correlations between fiber properties and carded ring yarn strength. Carded ring yarn strength Fiber quality properties Correlations UHML UI SFC FS FE Spearman (Rs) 0.97 0.36 - 0.57 0.98 - 0.29 40 Ne Probability < 0.0001 0.0993 0.0066 <0.0001 0.1946 Spearman partial (PRs) 0.51 0.43 - 0.09 0.43 - 0.01 Probability 0.0419 0.0959 0.7190 0.0941 0.9692 Spearman (Rs) 0.96 0.36 - 0.56 0.97 - 0.29 Probability < 0.0001 0.0993 0.0075 <0.0001 0.1946 50 Ne Spearman partial (PRs) 0.37 0.19 - 0.05 0.33 0.04 Probability 0.1512 0.4586 0.8415 0.2018 0.8766 respectively, according to [9], hence, these criterias are the main determinant for technological value of Egyptian cotton varieties. Generally, the extra long staple cotton varieties (Giza70, Giza87, Giza88 and Giza92) were the highest technological value than of long staple cotton varieties (Giza80 and Giza90), however, the long staple cotton variety Giza86 showed high technological value because of the increase in length and strength of the lint on those long staple cotton category to which it belongs. 3.2. Spearman (Rs) and spearman partial (PRs) correlations between criteria and the end product Spearman (Rs) and spearman partial (PRs) correlations coefficient of fiber quality properties (criteria), i.e., UHML, UI, SFC, FS, FE and FF used in modeling AHP with the end product represented in carded ring yarn strength at 40 and 50 count number are shown in Table 4. With yarn strength, the highest Spearman (Rs) correlations were of UHML and FS followed by UI and SFC. These variables showed spearman partial (PRs) correlation with yarn strength in the same trend, except for SFC, which exhibited quit low partial correlation. On the other hand, micronaire value (FF) that gave low value of spearman correlation, Unexpected showed partial correlation with yarn strength at count 40(Ne) higher than its spearman correlation. Whereas, partial correlation of micronaire value with yarn strength at count 50 (Ne) was lower than spearman correlation. This is due to the residual effect of yarn count that did not use in modeling AHP and resulted in shortcoming in micronaire value effect. Short fiber content (SFC) showed high significant correlation with yarn strength; whereas it is partial correlation with yarn strength was insignificant. This result clarifies the interrelation between SFC and the other variables with yarn strength. It is apparent that neither spearman correlation nor partial correlations of fiber elongation (FE) with yarn strength were insignificant at the two count numbers (Ne) 40 and 50, this suggests that, there no imperative use of FE in modeling AHP. Generally, data in Table 4 managed to verify the accuracy of weights of variables used to form AHP equation, except for fiber elongation (FE) which maybe released in other runs. 4. CONCLUSIONS By the two approaches of Saaty and Alonso and Lamata, we had acceptance consistency in pair-wise reciprocal comparison in the analytic hierarchy process (AHP) as a quality criterion determining the technological value of cotton. 4 FF 0.14 0.5197 0.33 0.2022 0.15 0.4990 0.14 0.6008 The two main determinant criterias for technological value were fiber length (UHML) and fiber strength (FS), which had the highest weights in MIAHP. Comparison between partial correlation and spearman correlation of fiber properties with yarn strength managed to verify the accuracy of weights of variables used to form MI AHP equation, except for fiber elongation (FE) which maybe released in other runs. 5. REFERENCES 1. T. L. Saaty, The Analytic Hierarchy Process. McGrawHill International, New York (1980). 2. T. L. Saaty, Highlights and critical points in the theory and application of the Analytic Hierarchy Process. European J. of Operational Res., 74 (1994) 426-447. 3. C. L. Hwang and K. Yoon, Multiple Attribute Decision Making: Methods and Applications. Springer-Verlag, New York, NY(1981). 4. T. L. Saaty, How to make a decision: The Analytic Hierarchy Process. European J. of Operational Res., 48 (1990) 9-26. 5. J. S. Dyer, Remarks on the Analytic Hierarchy Process. Management Sci., 36 (3) (1990) 249-258. 6. J. S. Dyer, A clarification of remarks on the Analytic Hierarchy Process. Management Sci., 36(3) (1990) 274-275. 7. A. Majumdar, B. Sarkar and P. K. Majumdar, Application of Analytic Hierarchy Process for the Selection of Cotton Fibers. Fibers and Polymers, Vol. 5 (4) (2004) 297-302. 8. A. Majumdar, P. K. Majumdar and B. Sarkar. Determination of the Technological Value of Cotton Fiber: A Comparative Study of the Traditional and Multiple- Criteria Decision-Making Approaches. Autex Research Journal, Vol. 5 (2) (2005) 71-80. 9. K.M. Hussein, A.A. Hassan and M.M. Kamal, The Multiplicative Analytic Hierarchy Process (MIAHP) as a Quality Criterion Determining the Technological Value of the Egyptian Cotton Varieties. Amer. J. of Plant Sci. 1(2010) 106-112. 10. K.M. Hussein and E. A. M. Ebaido, Comparison Of Quantification Methods Of Egyptian Cotton Fiber Quality. Bull. Fac. Agric., Cairo Univ., 62 (3) (2011) 285-292. 11. T. L. Saaty, Axiomatic foundation of the Analytic Hierarchy Process. Management Sci., 32 (7) (1983) 841-855. 12. J.A. Alonso and Teresa Lamata, Consistency in the analytic hierarchy process: A New Approach. International Journal of Fiber and Textile Research 2013; 3(1): 1-5 International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, 14 (4) (2006) 445−459. 13. ASTM, American Society for Testing and Materials. Designation, (D4605-86 - 1776-98) Test Method for Measurement of Cotton Fibers by High Volume Instruments (HVI) (1998). (D- 1440- 65). (D- 1578-67, 1998). Philadelphia 3, Pa, U.S.A. 14. J. Lin, A. Yang and A. Shah, Using SAS to Compute Partial Correlation. PharmaSUG2010 - Paper SP01. Merck & Co., Inc. Rahway, NJ 07065.USA (2010)1-5. 15. SAS. SAS version 9.1. User’s guide. Cary, NC: SAS institute, Inc (2004). Source of support: Nil; Conflict of interest: None declared 5 International Journal of Fiber and Textile Research 2013; 3(1): 1-5
© Copyright 2024 Paperzz