Hybrid Intelligent Optimal Algorithms With Applications

Simple in the concept, easy to be implemented and efficient, intelligent optimization algorithms, which use the information interchange and cooperation between individuals to achieve the optimization, have been successfully used in various real-world problems.And intelligence algorithms have become a hot research area.The improvements and applications have been discussed in this paper.Firstly, an improved artificial bee colony(ABC) algorithm (HABC) is presented to address the single objective optimal problems.ABC algorithm with few control parameters has been proposed recently, extremely efficient for solving multimodal and multidimensional optimization problems.The main contributions of our work are to enhance the search efficiency with the improved search mechanisms:(1) a stochastic gradient search is integrated into ABC algorithm, thereby enhancing the exploitation abilities of the employed bees; (2) a novel search mechanism with the idea of the differential evolution (DE) is used by onlookers, so as to improve the exploration abilities.The comparisons of numerical experiments among HABC, ABC, DE, particle swarm optimization algorithm are done, which show HABC algorithm can obtain better mean solution with faster convergent velocity and be more robust.To sum up, HABC algorithm outperforms other compared algorithms.Secondly, after studying a few classical evolution algorithms for the multiobjective optimization problems, a novel hybrid evolution algorithm (HMOGA) with the advantages of these classical algorithms is proposed.HMOGA not only defines the fitness function by the domination count and introduces a crowded-comparison approach to a spread of solutions, but also incorporates an archive truncation method to accelerate the convergence.With the increasing number of the nondominated solutions, k-nearest neighbors operator is been applied to save the nondominated solutions of larger distance.Numerical experiments show that HMOGA is able to come closer to the true front than classical evolutional algorithms.Finally, control and synchronization of chaotic systems are studied.At first, these problems are transformed into numerical optimization problems; Furthermore, a novel hybridization of differential evolution (DE) and bacterial foraging algorithm (BFOA), called CDEM algorithm, is proposed.CDEM algorithm not only incorporates an adaptive chemotactic step borrowed from the realm of BFOA into DE, which could improve the convergence characteristics of the classical DE, but also introduces mutation operation of genetic algorithm for enhancing population diversity.Lastly, CDEM algorithm is used to solve the problems of chaotic system control and synchronization.Numerical simulations based on Hénon Map demonstrate the effectiveness and stability of the CDEM algorithm, and the effects of some parameters are investigated as well.