Related Experiment Video
Updated: Jun 11, 2025
Spatial Multiobjective Optimization of Agricultural Conservation Practices using a SWAT Model and an Evolutionary Algorithm
Published on: December 9, 2012
An ensemble optimizer with a stacking ensemble surrogate model for identification of groundwater contamination source
Liuzhi Zhu1, Wenxi Lu1, Chengming Luo1
1Key Laboratory of Groundwater Resources and Environment, Ministry of Education, Jilin University, Changchun 130021, China; Jilin Provincial Key Laboratory of Water Resources and Environment, Jilin University, Changchun 130021, China; College of New Energy and Environment, Jilin University, Changchun 130021, China.
This study introduces an ensemble learning framework for groundwater contamination source identification (GCSI), significantly reducing simulation time and improving identification accuracy. The new method enhances efficiency and reliability in addressing complex environmental challenges.
Area of Science:
- Environmental Science
- Hydrogeology
- Computational Science
Background:
- Groundwater contamination source identification (GCSI) is crucial for environmental protection.
- Traditional simulation-optimization methods face challenges in time cost and accuracy due to problem complexity.
Purpose of the Study:
- To develop an innovative ensemble learning framework to overcome the limitations of traditional GCSI methods.
- To enhance both the efficiency and accuracy of groundwater contamination source identification.
Main Methods:
- Developed a stacking ensemble model (SEM) integrating Extremely Randomized Trees, Adaptive Boosting, and Bidirectional Gated Recurrent Unit as a surrogate for simulation.
- Created an ensemble optimizer (E-GKSEEFO) combining Genghis Khan Shark Optimizer and Electric Eel Foraging Optimizer for enhanced search strategies.
- Validated the SEM-E-GKSEEFO framework using hypothetical scenarios from a real coal gangue pile.
Main Results:
- The SEM demonstrated superior fitting performance over single machine learning models for high-dimensional, nonlinear GCSI data.
- The E-GKSEEFO significantly improved the accuracy of GCSI results compared to individual optimizers.
- The integrated SEM-E-GKSEEFO framework effectively reduced time costs while maintaining high identification accuracy.
Conclusions:
- The proposed SEM-E-GKSEEFO ensemble inversion framework is effective and superior for GCSI.
- This approach offers a promising solution for accurate and efficient identification of groundwater contamination sources.
