Softcomputing techniques are receiving attention as optimisation, techniques for many industrial applications. Although these techniques eliminate the need for derivatives computation, they require much work to adjust...
详细信息
Softcomputing techniques are receiving attention as optimisation, techniques for many industrial applications. Although these techniques eliminate the need for derivatives computation, they require much work to adjust their parameters at the stage of research and development. Issues Such as speed, stability, and parameters convergence remain much to be investigated. This paper discusses the application of the method of reference model to determine parameters of asynchronous machines using two optimisation techniques. Softcomputing techniques used in this paper are evolutionary strategy and the chemotaxis algorithm. Identification results using the two techniques arc presented and compared with respect to the conventional simplex technique of Nelder and Mead. Discussion about the chemotaxis algorithm as the most promising optimisation technique is presented, giving its advantages and disadvantages.
Purpose - The purpose of this paper is to apply a hybrid algorithm based on the combination of two heuristics inspired by artificial life to the solution of optimization problems. Design/methodology/approach - The flo...
详细信息
Purpose - The purpose of this paper is to apply a hybrid algorithm based on the combination of two heuristics inspired by artificial life to the solution of optimization problems. Design/methodology/approach - The flock-of-starlings optimization (FSO) and the bacterial chemotaxis algorithm (BCA) were adapted to implement a hybrid and parallel algorithm: the FSO has been powerfully employed for exploring the whole space of solutions, whereas the BCA has been used to refine the FSO-found solutions, thanks to its better performances in local search. Findings - A good solution of the 8-th parameters version of the TEAM problem 22 is obtained by using a maximum 200 FSO steps combined with 20 BCA steps. Tests on an analytical function are presented in order to compare FSO, PSO and FSO + BCA algorithms. Practical implications - The development of an efficient method for the solution of optimization problems, exploiting the different characteristic of the two heuristic approaches. Originality/value - The paper shows the combination and the interaction of stochastic methods having different exploration properties, which allows new algorithms able to produce effective solutions of multimodal optimization problems, with an acceptable computational cost, to be defined.
暂无评论