Neural ordinary differential grey algorithm to forecasting MEVW systems

Because of the advantage of the gray theory for forecasting small sample time data, gray algorithm theory has definitely been extensively utilized since it has been proposed and is currently being widely developed for predicting frames particularly in small sample problems. This article presented a viewpoint called gray algorithm by neuron- based ordinary-differential equation (NODE), called NODGM (neuron-based ordinary-differential gray-mode). In this type, we learn prediction methods through a training process that includes whiting equations. Compared with other models, the structure and time series via the regularity of real-samples are required in advance, so this NODGM design can have a better feasibility of applications and also study the origins of data according to different samples. The purpose is obtaining a better design with high forecast effectiveness, this study uses NODGM to train the model, while Runge-Kutta method is used to have the forecast set and solve numerical framwork. This algorithmic design creates a favorable theoretical basis for the installation of new process and distributes the numerical dimensions of completely mechanically elastic vehicle wheels (MEVW) in practical simulations.