Accuracy Optimization of Synchronization Parameters in Chaotic Time Series Prediction

Abstract

In view of the situation that current chaotic time series prediction model still exists the problem of low accuracy, this paper proposes a synchronous parameter optimization strategy of chaotic time series model based on improved genetic algorithm. It firstly takes the random direction method to construct the initial population of genetic algorithm to speed up the convergence, and then prescribes a limit to the objective function with improved punishment function. Then it optimizes the crossover, mutation and duplication operators of the genetic operator to avoid the premature convergence. Moreover, it adopts the complex method to optimize the local optimization ability of the standard genetic algorithm on the basis of improved fitness function. Finally, this paper applies the improved algorithm into the synchronous parameters optimization of chaotic time series prediction model. Simulation results show that the proposed algorithm has higher accuracy and efficiency in parameter optimization

Keywords

• Chaotic time series
• chaotic prediction
• synchronization parameters optimization
• improved genetic algorithm
• accuracy optimization
• genetic operator
• complex method
• punishment function optimization. ___________________________________________________________________