As pointed out above, the linear and analytic approximation enables the calculations of real systems operating with complex relations by applying the matrix method in an automated and computerised way. Any real system has to be represented by such complex serial and parallel connections which can hardly be reviewed, let alone calculated by humans.
In the course of this, by analogy with hypermatrices, summarizing and partitioning representations are built enabling the application of matrix operations to realise algorithms used in information technology. The first such representation was the matrix type form of the feedback control circuit which was extended so as to include the controllers and the disturbances affecting such systems. The terms calculating the various quality requirements and their norms are built in the controllers. Using the same technique, the full system representations developed this way get the possibility of fractional transformation which can ensure with the above mentioned partitional forms the stability of the entire system and its components, as well as the definition and control of further quality requirements built in the system.
In doing so, we have to bear in mind the specific limits of models (mainly those applying superposition) that substantially differ from reality.