Genetic algorithms assisted machine learning algorithms to optimize nano-phytoremediation of cadmium designed by...

Area of Science:

• Environmental Science
• Nanotechnology
• Biotechnology

Background:

• Hazardous contaminants like cadmium (Cd) pose significant environmental risks.
• Phytoremediation offers a sustainable approach for contaminant removal.
• Enhancing phytoremediation efficiency is crucial for effective environmental cleanup.

Purpose of the Study:

• To investigate the efficacy of titanium dioxide (TiO2) nanoparticles in enhancing cadmium (Cd) phytoremediation.
• To utilize artificial intelligence and machine learning for predicting and optimizing Cd absorption rates.
• To identify optimal conditions for Cd phytoremediation using computational modeling.

Main Methods:

• Experimental design with 20 combinations of Cd and TiO2 nanoparticle concentrations and exposure times.
• Response Surface Regression Analysis for identifying optimal input factors.
• Machine learning models (Gaussian Process Regressor) for predicting Cd absorption, evaluated using R^2 and MSE metrics.
• Genetic Algorithm (GA) for minimizing prediction errors and optimizing process parameters.

Main Results:

• TiO2 nanoparticles significantly increased Cd uptake by the plant.
• Ceratophyllum demersum achieved 99.58% Cd absorption, reducing Cd concentration to 0.0199 mg/L.
• The Gaussian Process Regressor model demonstrated high accuracy (R^2=0.99, MSE=0.07).
• The Genetic Algorithm identified optimal parameters for enhanced Cd absorption.

Conclusions:

• Computational models, particularly GPR, accurately predict Cd absorption in phytoremediation.
• A synergistic relationship between Cd concentration and treatment time enhances absorption.
• Supplementation with TiO2 nanoparticles significantly improves phytoremediation efficiency for cadmium removal.