A Diversity Model Based on Dimension Entropy and Its Application to Swarm Intelligence Algorithm

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Swarm Intelligence

Background:

• Swarm intelligence algorithms are valuable for optimization due to self-organization, self-adaptation, and self-learning.
• Traditional swarm intelligence algorithms suffer from premature convergence and loss of population diversity on high-dimensional, complex multi-peak problems.

Purpose of the Study:

• To address the premature convergence issue in swarm intelligence algorithms.
• To propose a novel approach for maintaining population diversity in complex optimization scenarios.

Main Methods:

• Introduced dimension entropy as a metric to quantify population diversity.
• Developed a diversity control mechanism to guide the swarm intelligence algorithm's updates.
• Implemented and tested the improved algorithm against traditional methods.

Main Results:

• The proposed dimension entropy effectively measures population diversity.
• The diversity control mechanism successfully maintains algorithm diversity in early stages and ensures convergence in later stages.
• Experimental results demonstrate superior performance of the improved algorithm over the original swarm intelligence algorithm.

Conclusions:

• The novel approach using dimension entropy and a diversity control mechanism effectively overcomes the premature convergence problem.
• This enhanced swarm intelligence algorithm shows improved performance for complex optimization tasks.
• The findings suggest a promising direction for developing more robust and efficient swarm intelligence algorithms.