Cooperative Particle Swarm Optimization With a Bilevel Resource Allocation Mechanism for Large-Scale Dynamic...

Area of Science:

• Optimization Algorithms
• Computational Intelligence
• Swarm Intelligence

Background:

• Cooperative coevolutionary algorithms struggle with dynamic optimization due to environmental changes, multiple fitness peaks, and subproblem unevenness.
• Existing resource allocation relies on past improvements, inefficiently using resources on hard subproblems or local optima.

Purpose of the Study:

• To propose a novel cooperative particle swarm optimization (PSO) algorithm for large-scale dynamic optimization.
• To address limitations in resource allocation for dynamic optimization problems.
• To enhance the adaptability and efficiency of cooperative coevolutionary algorithms.

Main Methods:

• Introduced a bilevel balanceable resource allocation mechanism within a cooperative PSO framework.
• Implemented a lower-level search strategy using solution diversity and quality to identify new peaks.
• Developed an upper-level resource allocation strategy balancing subproblem coevolution based on historical performance and potential improvements.

Main Results:

• The proposed algorithm demonstrates competitive performance against state-of-the-art methods.
• Achieved improved objective function values in large-scale dynamic optimization tasks.
• Showcased enhanced response efficiency to environmental changes.

Conclusions:

• The bilevel resource allocation mechanism effectively addresses resource allocation challenges in cooperative coevolutionary algorithms.
• The proposed cooperative PSO algorithm offers a promising approach for tackling complex dynamic optimization problems.
• The algorithm shows significant improvements in both solution quality and adaptability.