A novel hybrid method based on Cuckoo optimization algorithm and artificial neural network to forecast world's carbon...

Area of Science:

• Energy Economics
• Environmental Science
• Computational Intelligence

Background:

• Global energy consumption and CO2 emissions are critical factors in climate change.
• Traditional Artificial Neural Networks (ANNs) face limitations like slow training and local optima.
• Accurate forecasting is essential for effective climate change mitigation strategies.

Purpose of the Study:

• To develop an improved Artificial Neural Network (ANN) model optimized by the Cuckoo Optimization Algorithm (COA), termed COANN.
• To forecast future global energy consumption and CO2 emissions.
• To evaluate the effectiveness of the COANN model in predicting greenhouse gas emissions.

Main Methods:

• Utilized historical data on primary energy, oil, coal, and natural gas consumption from British Petroleum.
• Developed a hybrid model combining the Cuckoo Optimization Algorithm (COA) with Artificial Neural Networks (ANNs) to create the COANN.
• Assessed COANN performance using statistical metrics: Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Correlation Coefficient (CC).

Main Results:

• The COANN model demonstrated superior performance compared to traditional ANNs, overcoming issues like slow training and local optima.
• The model successfully forecasted CO2 emissions globally by 2050.
• Performance evaluation metrics confirmed the reliability and accuracy of the COANN model.

Conclusions:

• The COANN is a robust and dependable tool for monitoring global warming and forecasting CO2 emissions.
• This methodology provides valuable insights for policymakers, researchers, and experts involved in greenhouse gas management.
• The proposed COANN model can significantly influence global climate change policies and governmental interventions.