Many-Objective Evolutionary Algorithm Based on Spectral Clustering and Population Distribution

Numerous real-world applications involve complex many-objective optimization problems characterized by a large number of objective functions. Despite the development of various algorithms to address this kind of problems, these approaches often suffer from insufficient selection pressure, complex Pareto front shapes, and uneven distribution of solutions. To address this issue, this paper introduces a many-objective evolutionary algorithm based on spectral clustering and population distribution. We first utilize spectral clustering to classify the objective functions based on their correlations and aggregate them within the same class to reducing the complexity of the problem. Then, we propose a knowledge-guided decomposition based evolutionary algorithm to solve the simplified model. Finally, we conduct a series of comparative experiments to evaluate the effectiveness of the proposed method. The experimental results indicate that our method demonstrates excellent performance across various problems.