Efficiency Prediction and Optimization of the Pearl River Delta Logistics Industry Using an Integrated DEA Malmquist Model and Multi Objective Gradient Boosted Decision Trees

With the continuous growth of global logistics demand, improving logistics efficiency and optimizing resource allocation have become key issues in modern logistics systems. This paper proposes a novel approach that integrates the DEA-Malmquist model and MO-GBDT (Multi-Objective Gradient Boosting Decision Trees) to optimize logistics efficiency in the Pearl River Delta (PRD) region. First, the DEA-Malmquist model is used for both static and dynamic efficiency evaluation, revealing differences among cities in terms of resource allocation and efficiency improvement. Then, the MO-GBDT model is applied for multi-objective optimization to simultaneously improve logistics efficiency, reduce costs, and optimize resource allocation. Through empirical analysis of data from the PRD region from 2017 to 2023, the effectiveness of the integrated model in enhancing efficiency and optimizing resources is demonstrated. Experimental results show that the integrated model significantly outperforms traditional methods, especially in handling multi-objective optimization problems. The study provides a new logistics optimization solution for the PRD region and offers valuable insights for logistics systems optimization in other regions.