An adaptive hybrid algorithm based on particle swarm optimization and differential evolution for global optimization.

Area of Science:

• Computational intelligence
• Optimization algorithms
• Metaheuristic search

Background:

• Particle Swarm Optimization (PSO) and Differential Evolution (DE) are widely used population-based stochastic search techniques.
• Both PSO and DE are prone to converging to local optima, limiting their effectiveness in complex optimization tasks.
• Existing variants of PSO and DE often struggle with maintaining population diversity and escaping local optima.

Purpose of the Study:

• To develop a novel adaptive hybrid algorithm (HPSO-DE) that integrates the strengths of PSO and DE.
• To enhance the ability of optimization algorithms to escape local optima and maintain population diversity.
• To investigate the performance of the proposed HPSO-DE against established PSO and DE algorithms.

Main Methods:

• Formulation of a hybrid algorithm (HPSO-DE) by balancing parameters between PSO and DE.
• Implementation of an adaptive mutation strategy triggered when the population clusters around local optima.
• Comparative performance analysis of HPSO-DE against PSO, DE, and their variants on various optimization problems.

Main Results:

• The HPSO-DE algorithm demonstrates competitive performance compared to standard PSO, DE, and their variants.
• Adaptive mutation effectively helps the population to jump out of local optima.
• The hybrid approach successfully maintains population diversity throughout the optimization process.

Conclusions:

• The proposed HPSO-DE algorithm offers an effective solution for overcoming local optima in population-based optimization.
• The adaptive parameter balancing and mutation strategy contribute to improved search capabilities.
• HPSO-DE presents a promising alternative for complex optimization problems in scientific and engineering fields.