An Extended Cultural Algorithm

Computational Intelligence is a new research direction with the evolutionary computation, artificial neural networks and fuzzy systems theory. There is a class of simulated biological behavior of swarm intelligence, called swarm intelligence (Swarm Intelligence) in the field of computational intelligence. From the 20th century early 90's, it came a simulation of natural biological communities (Swarm) optimization behavior. Dorigo have been inspired from the mechanism of biological evolution simulating routing ant behavior of ant colony optimization method proposed (Ant Colony Optimization); Fberhart and Kennedy advanced Particle Swarm Optimization (Particle Swarm Optimization, referred to as PSO) in 1995 from the study of the birds predatory behavior, particle swarm optimization (PSO) was first to address the issue of continuous optimization; Reynolds advanced a cultural algorithm (Cultural Algorithm) in 1994. Cultural algorithm is advanced for computing systems in the evolution of growth over time to accumulate experience in the evolution of individual model components. These three algorithms are based on evolutionary algorithm groups and groups of individuals share information through social interaction.To improve the parallel Computational ability of the basic Cultural Algorithm (CA) and the calculation accuracy and computational efficiency of particle swarm optimization (PSO) algorithm. In this paper, the integrated use of the characteristics of two algorithms, An extended Cultural Algorithm (ECA) is proposed. It takes the particle swarm algorithm space into the framework of the culture algorithm, divided into several small groups of space, with the corresponding of a small number belief spaces controled by the fuzzy system consisting of inertia weight, they constitute the big belief space of cultural algorithm. Among them, the space for each small group has its own evolutionary way, and regularly contribute the good knowledge to the belief space, The belief space, through the evolution, take the influence by form of influence function to the following groups of space, So formate the "double the evolution of double promotion" mechanism of cultural algorithm.The main contributions of this paper are listed as follows:Advanced a extended cultural algorithm based on fuzzy adaptive PSO. It takes the particle swarm algorithm space into the population space of the culture algorithm and take the fuzzy control system into the belief space of cultural algorithm, extended to the multi-faith space between groups of "double-double to promote evolution", not only improve the calculation of particle swarm optimization validity, but also to enhance the cultural capacity of parallel computing algorithms. Simulations for a series of benchmark test functions show that the proposed method possess better ability to find the global optimum, especially, as the problem dimension scales up.