|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|83994||2018||28 صفحه PDF||سفارش دهید||11164 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Expert Systems with Applications, Volume 106, 15 September 2018, Pages 263-276
The selection of an optimal, efficient and reliable transport and logistics policy is one of the most important factors in supply chain management and logistics planning. Given that decision making in transport and logistics involves the consideration of a number of opposite criteria and possible solutions, such a selection can be considered as a multi-criteria decision making (MCDM) problem. This study presents a new integrated approach to decision making, i.e. the generation of a robust decision making rule (RDMR) by combining different MCDM methods and Taguchi's robust quality engineering principles. As a logical implication of the proposed approach, a conceptual model of an adaptive and interactive expert system is developed. Its purpose is to enhance the decision making process through enabling the decision maker to: (i) use higher level knowledge regarding the selection of criteria weights and MCDM methods, (ii) estimate the ranking of a new alternative, which can be added to the initial decision matrix after a posteriori analysis of the final rankings of alternatives, and, moreover, quantify its distance from the ideal and the anti-ideal solution. Five different case studies in the field of transport and logistics were considered in order to illustrate the proposed approach. The obtained results and the results of the other MCDM methods were compared using Kendall's tau-b and Spearman's rho tests. An analysis of the final rank stability with respect to the changes in criteria weights was also performed so as to assess the sensitivity of the alternative rankings obtained by the application of different MCDM methods and the proposed approach. To this aim, the Monte Carlo simulation was conducted covering 1000 different scenarios of criteria weights in three different cases. Finally, an additional procedure was introduced for an explicit representation of an RDMR using the design of experiments (DOE) principles.