Development Of Multi Attribute Decision Making Technique For Improved Performance In Manufacturing And Supply Chain Function

Abstract

Due to globalization decision making are very important if growth of product variation, reduction in product life cycle, splitting of supply chain, changing technologies and sustainability. Decision making is a part of operation research. When there are numbers of alternatives rank according to consider the complex criteria (beneficial/ non-beneficial and with different units), it is very difficult to make a transparent decision without favoritism to meet the objective. There are several techniques proposed by various researchers for solving through multi attribute decision making (MADM). A major criticism of MADMs is that different techniques may yield different ranking solution when applied to same problem; there are several limitations of current MADM with various mathematical set theories. Here the fundamental research is carried out, where the issues and limitations of current MADM techniques are tried to be resolved. The Single Valued Neutrosophic Set (SVNS) is an ideal set of Neutrosophic theory, which incorporates the data in level of truthness, level of indeterminacy and level of falsehood. Here, Three new approaches for MADMs are tried to investigate (i) Fuzzy Single Valued Neutrosophic Set Novel MADM(F-SVNS N- MADM), (ii) Fuzzy Single Valued Neutrosophic Set Entropy Weight Based MADMD (F-SVNS EW-MADM) and (iii) Fuzzy Single Valued Neutrosophic Set Advanced Correlation Coefficient MADM (F-SVNS ACC-MADM); which works with conversion on crisp/ fuzzy set into single valued Neutrosophic set. The proposed methodologies are implanted with same input information in some case examples published in various peer reviewed journals and books in the field of manufacturing and supply chain management. Validation is carried out with two stages (i) Comparison with published result of MADM and (ii) Spearman rank correlation coefficient sensitivity analysis of three methodologies with various normalization methods. The outcome of implementation shows that proposed methodologies give more accurate result with less calcula