Research on Strategic Supplier Selection and Evaluation Method of Manufacturing Enterprise Based on AHP-GCA Method

Abstract

With the application of intelligence technology in manufacturing industry, the competition of supply chain in manufacturing industry is becoming more and more competitive. Strategic supplier is a key element for the success of manufacturing enterprise. At present, there are many methods for the selection and evaluation of strategic supplier of manufacturing enterprise. Analytical hierarchy process and grey correlation analysis (AHP-GCA) are methods used to choose strategic supplier of manufacturing enterprise by building optimum reference data and solving grey correlation degree. These have positive effects to the enterprise practice in qualitative diagnosis, quantitative diagnosis and grey information processing. Introduction The standards of strategic supplier selection and evaluation system have obvious characteristic of hierarchy. The first level standards can be disassembled into different sub-indexes which form tree structure that offers structure basis for analytic hierarchy method. Analytic hierarchy method is a relatively mature theory which has much hands-on experience to draw on. Grey correlation analysis is a method which carries out quantitative description and comparison towards system development and change. Its basic idea is to judge whether the relation between reference sequence of number and comparison sequence of number is close or not by confirming the degree of closeness/ proximity of these two curves, and use grey correlation degree to reflect the proximity between curves and find out the difference and closeness. AHP-GCA (Analytic hierarchy processing and grey correlation analysis) model combine analytic hierarchy processing method with grey correlation analysis method organically. In AHP-GCA (Analytic hierarchy processing and grey correlation analysis) model, different indexes’ weight are confirmed by using analytic hierarchy processing method. Human’s subjective judgment is expressed in number, and consistency check is carried out, thus the side-effect of subjective factors could be brought down, and drastically. Grey correlation analysis method implements whitening processing on qualitative index items (grey indexes) from indexes, and standardizes quantitative index data. By this mean, the losing and deviation of information would be minimized and the correctness of conclusion could be guaranteed. Selecting Evaluation Model Based on AHP-GCA First step, building tree structure model. Second step, constructing “multiple judgment” matrix, k B signifies number k index in the B gradation. Suppose k’s indexes at this gradation is 1 B , ..., n B , its direct upper level index is A, then make multiple comparison to 1 B , ..., n B aiming at index A by using nine points marking method. Thus comparison value ij b is appeared. We note down judgment matrix (b ) ij n n B   . So B’s maximum eigenvalue is max  , normalized eigenvector that belongs to max  is 1 ( , , ) T n      . In this way, 1 ( , , ) n    is weighting of index 1 B , ..., n B to index A. In actual computation, we can calculate 2015 International Conference on Humanities and Social Science Research (ICHSSR 2015) © 2015. The authors Published by Atlantis Press 123 approximately index relative weighting under every single index by using “extraction of root” method. And after normalization, it is relatively important weighting of indexes at the same gradation to indexes at the direct upper gradation. In the similar way, we calculate from upper gradation to lower gradation until it comes to the last gradation which has its weighting to the direct upper gradation. If we note k D as matrix organized by weighting in column of all indexes in number k gradation to all indexes in the direct upper gradation, then the combined weighting vector in gradation is: 1 2 1 k k k W D D D D      . (1) While getting judgment matrix B, index consistency C I  is needed check due to probable existence of subjective judgment inconsistency of nine points marking method.