About the limits of applicability of the analytical network process in the decision-making problems in the natural sciences

The article reviews the concept, features, some of the key points and drawbacks of the analytical network process (ANP) for decision-making problems. It examines one of the key issues of the ANP – the problem of cycles and methods of its decision which are based on the experimental data. On the basis of these ones is formed the author’s view of the networks classification. It is inferred about the applicability of ANP for different types of networks for solution of scientific and some other kinds of applied tasks. For datasets that are unsuitable for processing by ANP is offered a different path, which consists in using the method of direct influences, developed by a group of scientists with participation of the author on the basis of the analytical networks concept as ANP alternatives. The offered path is also used for tasks that do not fall within the scope of ANP.