Adaptive Differentiated Parrot Optimization: A Multi-Strategy Enhanced Algorithm for Global Optimization with Wind...

Area of Science:

• Computational Intelligence
• Metaheuristic Optimization
• Nature-Inspired Algorithms

Background:

• The Parrot Optimization Algorithm (PO) is a nature-inspired metaheuristic based on parrot behavior.
• PO faces challenges with population diversity and early convergence in complex optimization problems.
• Existing algorithms struggle to maintain search effectiveness and identify optimal solutions.

Purpose of the Study:

• To introduce the Adaptive Differentiated Parrot Optimization Algorithm (ADPO) to overcome PO's limitations.
• To enhance exploration, exploitation, and convergence capabilities of the Parrot Optimization Algorithm.
• To validate ADPO's effectiveness on benchmark functions and a real-world application.

Main Methods:

• Developed ADPO with three novel mechanisms: Mean Differential Variation (MDV), Dimension Learning-Based Hunting (DLH), and Enhanced Adaptive Mutualism (EAM).
• MDV employs dual-phase mutation for balanced exploration and exploitation.
• DLH uses dimension-wise learning to prevent premature convergence and maintain diversity.
• EAM introduces fitness-guided interactions for balanced intensification and diversification.

Main Results:

• ADPO demonstrated superior convergence speed, search efficiency, and solution precision on CEC2017 and CEC2022 benchmark functions.
• In wind power forecasting with LSTM, ADPO achieved an average R² of 0.9726, outperforming conventional methods.
• ADPO consistently achieved superior Friedman rankings (1.42-2.78) against 12 advanced algorithms.

Conclusions:

• The proposed ADPO significantly enhances optimization capabilities compared to the baseline PO.
• ADPO shows robust performance and effectiveness in complex optimization and renewable energy prediction.
• The novel mechanisms effectively address population diversity and convergence issues in metaheuristic search.