A Multimodal Genetic Algorithm based Parameter Optimization Method for Flight Control System

Many engineers face challenges in parameter design and adjustment for complex systems with internal implicit relationships, particularly in flight control systems. Traditional parameter adjustment methods often struggle to fully consider the system's pre-constraints. This paper proposes a multimodal parameter design algorithm aimed at exploring the solution space comprehensively. It addresses the non-convexity of the solution space in the parameter design process of carrier-based aircraft flight control systems, as well as the challenge of reaching the global optimum due to multiple constraints. The algorithm leverages the mutual attraction and repulsion characteristics of genetic algorithm evolution. Additionally, it utilizes niche technology based on the K-means clustering algorithm, enabling multiple populations to search their respective spaces and obtain feasible solutions in different regions of the solution space. This approach facilitates exploration of feasible parameter regions and their relationships within the flight system, while also providing insights into the characteristics of the solution space.