|
Warehouse selection: a hybrid fuzzy multi-criteria decision-making approach |
|---|---|
| รหัสดีโอไอ | |
| Title | Warehouse selection: a hybrid fuzzy multi-criteria decision-making approach |
| Creator | Nay Chi Moe Oo |
| Contributor | Pham Duc Tai, Advisor |
| Publisher | Thammasat University |
| Publication Year | 2568 |
| Keyword | Warehouse selection, FBWM, TOPSIS, MCDM, Normalization techniques, Distance metrics, Sensitivity analysis, Robustness analysis, Two-way ANOVA |
| Abstract | Warehouse location selection requires the consideration of multiple, often conflicting criteria such as cost, space availability, and accessibility, as the warehouse itself plays a critical role in optimizing logistics costs and enhancing customer service. To accommodate the selection efforts, this study presents an integrated fuzzy multi-criteria decision-making approach that combines the Fuzzy Best-Worst Method (FBWM) with the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) to identify the most suitable warehouse location. The former is employed to determine the relative weights of criteria, taking into account the uncertainty inherent in expert judgments, while the later is used to rank the alternative locations with respective to the criteria and its weights.A case study, which involves three warehouse alternatives evaluated based on area, rental rate, and distance to the airport is conducted to demonstrate the effectiveness of the proposed method. Closeness coefficients were calculated across multiple methodological configurations using three normalization techniques (linear vector, linear sum, and max) and two distance metrics (Euclidean and Manhattan). To further explore the robustness of the rankings, combinations of weight generated using a complementary weighting strategy was experimented. From a discrete set of weights ranging from 0.05 to 0.90, a total of 5,832 possible combinations were generated. Two filtering conditions were applied to eliminate invalid weight combinations: all the weights must sum to one, and the weight for the rental cost criterion must be the largest one. This process yielded 45 valid weight combinations. These combinations of weight were later put into usage to evaluate the consistency of ranking outcomes.Sensitivity and robustness analyses reveal that the top-ranked warehouse (Alternative S2) consistently outperforms others regardless of methodological configurations and weights combinations. This confirms the reliability of the decision. In addition, Analysis of variance (ANOVA) results indicate that both weight combinations and distance metrics significantly affect the closeness coefficient (〖CC〗_i), while the normalization method shows minimal impact. Moreover, Manhattan distance provides higher discrimination among alternatives, whereas Euclidean distance offers more stable and consistent rankings. Overall, the proposed approach is robust and practical, providing decision-makers with a clear and reliable framework for selecting warehouse locations. |
