Integrated Assessment of the Effectiveness and Structuring of Objects in the Internet of Things System

The paper deals with the construction of Internet of things systems. The solution of two problems is considered. The first problem is to build an integrated assessment of system objects effectiveness, based on analysis the information of indicators achieved by objects characteristics values. Choice of rational variant of objects structuring and their grouping in the appropriate way depends on the choice of integral estimation model variant. To solve the grouping problem the approach based on neural networks is used. The parameters of the first neural network are learning rate, normalization parameter, and similarity criterion. In second neural network, this network uses a reset mechanism-temporary deactivation of the group if the similarity condition is not met. Ordering the numbering of Internet of things objects system by value of integral assessment of their functioning efficiency and grouping them into a given number of ordered groups leads to a certain variant of system structuring. Improving the efficiency of an integrated structured system is achieved through the use of transformational mechanisms. The first of them is objects absorption of the leading groups of objects from the groups of outsiders and affects the structural efficiency of the Internet of things, as it leads of the original set of objects reduction. The second of them ensures movement of objects in the leading groups through the rational use of the the resource allocated by the management center and affects the resource efficiency of the entire system. The optimization of mechanisms modeling leads to a number of discrete programming problems. Numerical methods for finding the best option for the transformation of the original system of the Internet of things, represented as a set of groups of objects requires on the one hand orientation to known approaches, and on the other taking into account the features of mathematical descriptions of transformation conditions and the possibility of constructing combined algorithms.