Chance-constrained multi-objective optimization of groundwater remediation design at DNAPLs-contaminated sites using...

Area of Science:

• Environmental Engineering
• Computational Science
• Optimization Theory

Background:

• Groundwater contamination by dense non-aqueous phase liquids (DNAPLs) poses significant remediation challenges.
• Optimizing remediation strategies requires balancing cost and time objectives.
• Efficient simulation models are crucial for complex environmental engineering problems.

Purpose of the Study:

• To propose and evaluate the multi-algorithm genetically adaptive multi-objective (AMALGAM) method for groundwater remediation design.
• To compare AMALGAM's performance against the non-dominated sorting genetic algorithm II (NSGA-II).
• To assess the accuracy of multi-gene genetic programming (MGGP) as a surrogate model for remediation simulations.

Main Methods:

• Implementation of the AMALGAM multi-objective optimization solver.
• Comparison with NSGA-II using remediation cost and time as objectives.
• Development of surrogate models using multi-gene genetic programming (MGGP), support vector regression (SVR), and Kriging (KRG).
• Incorporation of surrogate modeling uncertainty using chance-constrained programming (CCP).

Main Results:

• AMALGAM achieved lower remediation costs and shorter remediation times than NSGA-II.
• The MGGP surrogate model demonstrated higher accuracy compared to SVR and Kriging.
• Remediation cost and time were found to increase with higher confidence levels.

Conclusions:

• AMALGAM is a superior optimization method for DNAPL groundwater remediation compared to NSGA-II.
• MGGP is an effective surrogate modeling technique for accelerating complex environmental simulations.
• Chance-constrained programming effectively integrates uncertainty into the optimization process, aiding decision-making.