Step-5: Classification of product families: The DMs chose the group with 17 product families because the commonality was adequate, the product families formed were compatible with the grouping logic of the company and the results were applicable. The commonality of the group with 17 product families was found to be 0.695, while its efficiency was 4.09%. At this stage when 17 product families were obtained, 7 product families (19 products) were assigned to the "innovative" family, while 1 product family (1 product) was assigned to the "functional" product family. The remaining 9 product families containing 28 products were identified to be hybrid product family, which was the subject-matter of this study. When each of the product families is
…show more content…
The criteria that could be used for the selection of DP location (Table 1) were discussed with the expert group, 11 criteria were identified including predictability of demand (C11), delivery reliability (C12), delivery lead time (C13), order size (C14), order frequency (C15), modular product design (C21), holding cost (C22), backorder cost (C23), production lead time (C31), process and human resources flexibility (C32) and the location of the bottleneck (C33).
Step-8: Identification of interactions and causal relationship between criteria by fuzzy Dematel: At this stage, the interactions and relationships between the criteria were evaluated. The method of converting fuzzy data into crisp scores (CFCS) was used for defuzzification. The steps of CFCS method were given in Appendix A. This method was preferred because it gave greater membership function and greater precise values. As a result of this method, direct-relation matrix Z was found. Then the steps of Dematel method were applied (see Appendix …show more content…
The sum of rows and columns of the T matrix regarding the criteria is presented in Table 4. In order to keep the complexity of the model at a manageable level in systems with a high number of criteria and to explain the structural relationship between the criteria, the threshold value p was used with a view to filtering the negligible effects in the T matrix. The average of all values in the T matrix was calculated (Sumrit and Anuntavoranich, 2013) and the threshold value p was found to be 0.339. The criteria taht are higher than the threshold and have an interaction between were indicated in bold in the T matrix.
Based on the Di - Rj values indicated in Table 4, the criteria were assigned into "cause" and "effect" groups. The cause group included the criteria C11, C12, C13, C21, C22 and C23 with positive Di - Rj value, and the effect group included the criteria C14, C15, C31, C32 and C33 with the negative Di - Rj value. As the effects of the criteria in the cause group were greater on the system, these criteria should be followed carefully. The criteria contained in the effect group were influenced by the criteria in the cause group and they could improve the system performance through modifications and improvements in the criteria influencing
The criteria that could be used for the selection of DP location (Table 1) were discussed with the expert group, 11 criteria were identified including predictability of demand (C11), delivery reliability (C12), delivery lead time (C13), order size (C14), order frequency (C15), modular product design (C21), holding cost (C22), backorder cost (C23), production lead time (C31), process and human resources flexibility (C32) and the location of the bottleneck (C33).
Step-8: Identification of interactions and causal relationship between criteria by fuzzy Dematel: At this stage, the interactions and relationships between the criteria were evaluated. The method of converting fuzzy data into crisp scores (CFCS) was used for defuzzification. The steps of CFCS method were given in Appendix A. This method was preferred because it gave greater membership function and greater precise values. As a result of this method, direct-relation matrix Z was found. Then the steps of Dematel method were applied (see Appendix …show more content…
The sum of rows and columns of the T matrix regarding the criteria is presented in Table 4. In order to keep the complexity of the model at a manageable level in systems with a high number of criteria and to explain the structural relationship between the criteria, the threshold value p was used with a view to filtering the negligible effects in the T matrix. The average of all values in the T matrix was calculated (Sumrit and Anuntavoranich, 2013) and the threshold value p was found to be 0.339. The criteria taht are higher than the threshold and have an interaction between were indicated in bold in the T matrix.
Based on the Di - Rj values indicated in Table 4, the criteria were assigned into "cause" and "effect" groups. The cause group included the criteria C11, C12, C13, C21, C22 and C23 with positive Di - Rj value, and the effect group included the criteria C14, C15, C31, C32 and C33 with the negative Di - Rj value. As the effects of the criteria in the cause group were greater on the system, these criteria should be followed carefully. The criteria contained in the effect group were influenced by the criteria in the cause group and they could improve the system performance through modifications and improvements in the criteria influencing