Selection of appropriate supplier(s) for success of an organization is particularly a valuable necessity, hence apart from the common criteria such as logistics, service and quality, this paper discusses most of the key decision variables which can play a critical role in case of the supplier selection. In this study, analytic network process (ANP) method is used because it considers the relationship between the criteria themselves; criteria and alternatives. Pair wise comparison between the model elements is necessary in ANP method. However, the decision makers make their judgments in fuzzy environment and prefer to use linguistic variables with number interval instead of crisp number for stating judgments. For these reasons, a fuzzy set is required to give an answer for the uncertainty. In fuzzy ANP model, experts have been making fuzzy pair wise comparisons; however, the importance of compared criteria or their priority may be different. In such a case, the judgment of expert regarding pair wise comparisons of elements can change. The new evaluations of experts should be obtained. Getting the evaluation of experts in each case may delay decision making. To overcome this difficulty, data related to fuzzy pair wise comparisons that reflect expert opinion is used in different artificial neural network (ANN) models for training. There is no need to consult the experts in ANN comparison matrix values due to learning feature of ANN. Another superiority of ANN model is that the weights search by pair wise comparison matrix can be found by ANN without a need for fuzzy extent analysis method. This research results thus indicate that the supplier selection process appears to be the most significant variable in deciding the success of the supply chain. Therefore, supplier selection should be done according to many different qualitative and quantitative criteria.
Key words: Supplier selection, fuzzy analytic network process (FANP), artificial neural network (ANN)
Copyright © 2020 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0