A Multipopulation Dynamic Adaptive Coevolutionary Strategy for Large-Scale Complex Optimization Problems

Area of Science:

• Computational intelligence
• Optimization algorithms
• Swarm intelligence

Background:

• Large-scale optimization problems present significant computational challenges.
• Existing methods often struggle with dynamic adaptation to problem characteristics.
• The need for enhanced global and local search capabilities is critical.

Purpose of the Study:

• To propose a multipopulation dynamic adaptive coevolutionary strategy.
• To dynamically and adaptively adjust connections between population particles.
• To improve optimization accuracy and convergence rates for large-scale problems.

Main Methods:

• Development of a dynamic adaptive evolutionary network (DAEN) model.
• Analysis of network evolution characteristics in collaborative particle search.
• Adaptive evolution of collaborative topology based on coupling connection strength and swarm type.

Main Results:

• The proposed algorithm demonstrates high optimization accuracy for high-dimensional and large-scale problems.
• Significant improvements in converging rate were observed.
• Enhanced global and local searching capabilities were achieved through adaptive topology evolution.

Conclusions:

• The multipopulation dynamic adaptive coevolutionary strategy effectively addresses large-scale optimization challenges.
• The DAEN model provides a robust framework for adaptive coevolution.
• The algorithm offers a promising solution for complex optimization environments.