Method for solving a non linear problem by iteration

摘要

In order to solve a non-linear problem, an iteration method is used in which, in each iteration step (11), a prediction result of the non-linear is solved by a plurality of prediction methods (15), to obtain a plurality of parallel predicted solutions. Then the optimum one of those predicted solutions is selected (16) and the corresponding prediction result is then the prediction result used in the next iteration step (11). Thus, at each iteration step, the optimum method is used from a plurality of prediction methods. Hence, the predicted solutions converge rapidly to a final solution. The optimum solution is determined by comparing the predicted solution of each prediction method with the prediction result of the previous iteration, e.g. on the basis of difference, absolute difference, or ratio.