Multimodal multi-objective evolutionary algorithm assisted by graph neural networks based selection mechanism

Key Points

• Enhanced performance is achieved through the implementation of a graph neural networks-based selection mechanism, with significant improvements noted.
• The algorithm effectively optimizes multiple objectives simultaneously, addressing different problem landscapes.
• This approach utilizes a multimodal evolutionary algorithm to identify and preserve diverse solutions, improving adaptability.
• The findings suggest that integrating graph neural networks can provide a stronger framework for selection processes in evolutionary computation.