Application of Genetic Algorithm in Meshless Optimization of Elastic Foundation with Ribbed Plates and Beams

Abstract

In order to overcome the mesh dependency of finite element method, the author proposes the application of genetic algorithm in meshless optimization of elastic foundation with ribbed plates and beams. The ribbed plate is regarded as a combination of plates and beams. Based on the meshless method and combined with genetic algorithm, the rib arrangement position of the rectangular ribbed plate is optimized to minimize the deflection of the center point of the ribbed plate under lateral load. Compared to traditional finite element methods, using the author's meshless method for rib position optimization analysis of ribbed plates does not require mesh reconstruction, and the nodes discretized on the plate and ribs always do not need to be changed. The results indicate that the deflection values of the center points corresponding to the second generation individuals are more concentrated, and there are also many individuals with smaller deflection values compared to the first generation. The hybrid genetic algorithm is indeed effective. The author added the constrained random direction method to form a hybrid genetic algorithm based on genetic algorithm, which accelerates convergence speed, reduces computational repetition rate, and significantly reduces the computational algebra of genetic algorithm to two or three generations, resulting in better results.