Research and application of function optimization based on artificial fish swarm algorithm

Abstract

The software reliability modeling is an important field in software reliability engineering. As the existing software reliability models are nonlinear, the parameters of these models are difficult to estimate. The artificial fish swarm algorithm is simple and can quickly jump out of local extremum. Now it has been applied to the parameter estimation. On the basis of the basic artificial fish swarm algorithm, this paper improves the algorithm to improve the speed of convergence and gain a strong ability to overcome the local extreme value because the improved algorithm ignores the crowded degree factor; moreover, we make the artificial fishes only to execute the preying behavior and moving behavior in the later stage of algorithm to reduce the visual field of artificial fishes through the introduction of the attenuation factor and thus to improve the precision. The results of simulation experiments verify the improved algorithm has the ideal rate of convergence and precision of optimization.