SVB: Senkung der Herstellungskosten mit nichtlinearen Modellen

A study was conducted to demonstrate how systematic experimentation followed by the development of nonlinear models resulted in significantly better recipes that were more cost-efficient. The study demonstrated that nonlinear modeling could be carried out using a variety of methods to achieve these objectives. The newer methods included feed-forward neural networks, kernel regression, and multivariate splines, which do not require a priori knowledge of the nonlinearities in the relations. There were many different types of neural networks, and some of them had practical uses in process industries. The first stage of the work was to optimize the aggregate mix according to the theory and algorithm of a given equation to achieve the results.