Degeneration Recognizing Clonal Selection Algorithm for Multimodal Optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Engineering Applications

Background:

• Clonal selection algorithm (CSA) is a population-based metaheuristic inspired by biological immunity.
• Complex engineering multimodal optimization problems often require significant computational resources.
• Existing CSA methods may neglect valuable information contained within eliminated solutions.

Purpose of the Study:

• To propose a novel computing speed improvement for the clonal selection algorithm (CSA).
• To introduce the degeneration recognizing clonal selection algorithm (DR-CSA) for complex engineering multimodal optimization.
• To reduce the computational cost of optimization by avoiding unnecessary evaluation operations.

Main Methods:

• Developed a degeneration recognizing (DR) method integrated into the CSA framework.
• DR-CSA identifies and pre-eliminates degenerated solutions within the population before evaluation.
• Utilizes knowledge from previously eliminated solutions to inform the identification of degenerated new populations.

Main Results:

• DR-CSA demonstrated effectiveness in reducing computational time for complex optimization problems.
• The proposed method achieved comparable accuracy to the standard CSA.
• Significant savings in evaluation times were observed, particularly in computationally intensive tasks.

Conclusions:

• The degeneration recognizing clonal selection algorithm (DR-CSA) offers a viable approach to accelerate optimization processes.
• This method efficiently leverages information from non-optimal areas to improve computational performance.
• DR-CSA is effective for solving real-world engineering optimization problems, such as the wet spinning coagulating process.