| Fruit Fly Optimization Algorithm(FOA)has been widely used in different areas of life.It was applied in financial field in this dissertation.Although FOA itself have certain ability of global optimization,it is easy to appear problem of premature convergence and lower accuracy in the certain of the optimization process.Therefore this dissertation proposes a Adaptive Step Fruit Fly Optimization Algorithm(AS-FOA).On the basis of analyzing characteristics of fruit fly flight and location,the dissertation introduces two concepts which are Stability threshold and Fitness step length factor simultaneously.The second,after the algorithm sets twice the fruit fly random direction,judge whether the optimization achieve stability threshold conditions.At the same time ajust the searching step length according to the fitness of fruit flies dynamically.Using the improved flies algorithm optimizing the coefficient of logistic and the spread value of GRNN network.The last,through the experimental analysis,the dissertation use the two models was applied to enterprise financial warning.We draw the following conclusion that the enterprise financial warning analysis of the test results.When improved algorithm optimize the logistic coefficient and GRNN parameters,it further improving the prediction accuracy,speeding up the convergence speed,and effectively overcome the interference of local optimal value.This algorithm has better global optimization performance. |
