基于分形理论与模拟退火算法的安全投资预测

在研究安全投资规律及影响因素的复杂性、不确定性的基础上,提出基于分形理论与模拟退火算法的安全投资预测新方法。该方法首先运用重标极差法分析了安全投资时间序列预测的可行性,然后利用模拟退火算法并结合对原始数据的一系列变换求解满足预测误差精度的平移量,得到安全投资时间序列的分形维数,最后根据分形统计模型可以预测出下一时间单位的安全投资。通过实例证明该方法的可行性和有效性,提高了决策过程的科学性,对企业制定未来安全投资总量具有借鉴作用。

Safety Investment Forecasting Based on Fractal Theory and Simulated Annealing Algorithm

Based on the study of law of safety investment and the complication and uncertainty of its influencing factors,a new forecasting method based on fractal theory and Simulated Annealing Algorithm is put forward.This method firstly analyzes the feasibility of the forecast of safety investment time series with rescaled range analysis method,and then uses Simulated Annealing Algorithm combined with a series of transformations of original data to solve motion value to meet the accuracy of forecast error,thus obtains the fractal dimension of safety investment time series,finally safety investment of next time unit can be forecasted based on fractal model.The application of the method to an example demonstrates that it is feasible and effective,and improves the scientificity of decision-making.This study provides reference for enterprise to decide the quantity of future safety investment.