A Hybrid Algorithm Of GSO And PSO Based On Cultural Framework

In recent years, Large-scale chemical industries have witnessed vigorous development; chemical industry process models become increasingly complex, the optimization problem of the model also tend to high dimension and multimodal. Swarm intelligence which comes from swarming behaviors of animal groups, has important theory significance and practical value in solving the practical optimization problems. At present, many researches are focused on several intelligent optimization algorithms:particle swarm algorithm, group search optimizer, ant colony algorithm, shuffled frog leaping algorithm, artificial bee colony algorithm and so on. This paper firstly concludes the theories and advantages of these algorithms, Each algorithm has its own characteristics and advantages, at the same time have their own defects and shortcomings. Particle Swarm Optimization (PSO) has a good global search ability, It’s good at finding a relatively optimal local search area in a large search space, Group Search Optimizer (GSO) has a good local search ability, it’s good at searching in a relatively optimal local search area. But with the development of Practical optimization problems which become more and more complicated, particle swarm optimization and group search optimizer Algorithms in solving large-scale complex problem is easy to fall into local optimal solution and premature convergence.Therefore, in this paper, aiming at these problems, give full play to the advantages of particle swarm optimization and group search optimizer algorithms in global search and local search ability, combining these two kinds of algorithms into the cultural algorithm framework, A Hybrid algorithm of GSO and PSO Based on cultural framework is presented in this paper. The algorithm framework consists of population space and belief space, The mechanism evolution of population space is according to PSO algorithm, and GSO algorithm is used to the mechanism evolution of belief space, these two kinds of space are through accept function and influence function associated together, Also the strategy of dynamic selection is introduced into population space to improve the convergence efficiency. And the parameter settings of the algorithm are analyzed in this part.Finally, the optimization of the6typical teat functions is experimented by CBGSPSO and other intelligent optimization algorithms, and later CBGSPSO is applied to the chemical optimization process problems. The effectiveness and reliability of the improved algorithm is confirmed by the calculated results.