对制约系统运动的含时间微商项的方程,引进记忆函数后可导得一个差分一积分方程,它能包容含初值的多个往值。构建和求解系统的差分一积分方程的原理,称为自忆性原理,相应导得的数学模型称为自忆性模型。自忆性模型现已应用到气象、海洋、水文、市场、农业、交通和能源等多个领域的建模、计算和预报,尤其在预报方面取得了显著成果,提高了预报准确率。文中还论述了自忆性一界门模型,讨论了自忆性模型应用效果好的原因,给出了气候预测的实例。
Setting a memory function for a governing equation of a system with a time derivation, a difference - integral equation, covering multi - initial values, can be deducted. The principle of setting up and solving the difference - integral equation is called as a self - memorization principle. The model based on the principle is different from other time series models, which are based on a random difference equation. The self - memorization model has been applied to meteorology, oceanography, hydrology, marketing, agriculture, transportation and energy and so on for modeling, computing and forecasting. In particular, it is successful for forecast issue pertinent to advancing the forecast accuracy. Besides, the self - memory - gate model is briefed, and the reason why the self- memo- rization model can improve the forecast is discussed as well. As a example, a climate prediction is presented.