Deep Surrogate Network Enhancement of Non-dominated Sorting Genetic Algorithms for Improved Multi-objective Optimization

This research paper introduces an innovative approach to multi-objective optimization by integrating a Deep Surrogate Network (DSN) with the Non-Dominated Sorting Genetic Algorithm (NSGA), resulting in the novel NSGA-DSN model. Aimed at addressing the computational challenges associated with traditional NSGA models, the NSGA-DSN model leverages the advanced predictive capabilities of deep learning to approximate the fitness landscape, thereby significantly reducing the number of fitness evaluations required during the optimization process. Our comparative analyses reveal that the NSGA-DSN model not only achieves faster convergence rates but also produces solutions of higher accuracy and reduced mean squared errors compared to conventional NSGA approaches. These advancements underscore the model’s potential to enhance the efficiency and effectiveness of solving complex multi-objective optimization problems. Despite its reliance on high-quality training data and the computational demands of deep learning, the NSGA-DSN model represents a significant step forward in evolutionary algorithms, offering a promising solution for real-world applications across various domains. This paper discusses the development, implementation, and comparative performance of the NSGA-DSN model, highlighting its contributions to the field of computational optimization.