|
A hybrid machine learning and multi-criteria decision-making approach for optimizing 3PL selection in sustainable supply chains |
|---|---|
| รหัสดีโอไอ | |
| Title | A hybrid machine learning and multi-criteria decision-making approach for optimizing 3PL selection in sustainable supply chains |
| Creator | Nu Hoang Tran Le |
| Contributor | Rujira Chaysiri, Advisor |
| Publisher | Thammasat University |
| Publication Year | 2568 |
| Keyword | Third-party logistics, Multi-criteria decision-making, Sustainable supply chains, CRITIC, MARCOS, Stepwise regression |
| Abstract | This study proposes an integrated evaluation framework that combines CRITIC, MARCOS, and Stepwise Regression to assess the performance of 20 third-party logistics (3PL) providers from 2020 to 2024. CRITIC derived weights reveal the persistent dominance of economic criteria, supported by service level and capability, and social and environmental dimensions, underscoring the importance of financial resilience, operational reliability, and global network coverage. MARCOS rankings show a widening performance gap between top-tier and lagging firms, suggesting structural disparities in asset capacity, service consistency, and sustainability readiness. Sensitivity analysis highlights geographical coverage, international freight volume, and sustainability indicators as the most influential factors shaping ranking variability. Stepwise Regression further validates these findings by identifying a recurring subset of predictors primarily equity ratio, geographical coverage, revenue, and contract logistics that consistently explain next-year performance, with multi-year datasets enhancing model stability. External validation using key performance indicators confirms strong alignment between the proposed framework and real-world outcomes. Collectively, the results demonstrate that 3PL competitiveness is driven primarily by economic and operational characteristics, while sustainability gains importance over time. The integrated framework offers a robust and practical tool for benchmarking provider performance and guiding managers in resource allocation, capability development, and long-term strategic planning. |
