Improved Heuristic-Particle-Swarm Optimizer And Its Application To The Steel Structural Optimal Design

With the quick advancing of structural optimization,traditional optimization algorithms always have convergence issues in multi-objective and multi-constraint,multi-variable,multi-extrema nonlinear problem,such as slow convergence,poor stability,easy to fall into local optimal solution,even fail to converge etc.However,intelligent optimization algorithms,with its more reliable,practical,scientific and reasonable advance,are taken attention by numerous structural optimization researchers.At present,more efficient,stable,and practical intelligent algorithms keep being proposed,and the application of structure optimization has changed from the past size optimization of simple single condition to size optimization of complex multi condition,even expanding toward the shape and topology optimization.This paper takes Heuristic Particle Swarm Optimizer(HPSO)as searching primary motor,and introducing Pareto optimal theory to realize the multi-objective optimization of HPSO.The subsequent work is adding the practical operators of Genetic Algorithm(GA)to improve the shortcoming of Particle Swarm Optimizer(PSO),which is easy to fall into local optimum,finally proposing a new Multi-objective HPSO-GA Hybrid Algorithm(MHGH).MHGH unions highlights from HPSO and GA,also consolidating some generally utilized multi-objective methodologies,for example,crowding-distance mechanism,the adaptive grid technique and tabu search.Results demonstrate that the algorithm in the complex optimization,such as frame and shell structures,has better convergence rate,high computational efficiency and good stability.In the seismic optimum design,calling SAP2000 by using algorithms codes in MATLAB to fulfill finite element analysis,including Pushover analysis and time history analysis,for the plane structures and large span space structures.At last,the computational consequence of the optimization algorithm proposed in this paper,by comparing with a variety of classical optimization algorithms,we can find that the HPSO-based multi-objective algorithm and its improved hybrid algorithm have a wider Pareto front,besides convergence precision,speed and stability also has increased significantly,the data provides a dependable and profitable reference for the future study of the structural optimization.