The report is focused on the current state in the field of structural analysis of systems. Most general approaches to defining structural similarity and methods for graph models structural similarity analysis are discussed. Implementation of the discussed methods in the various subsystems of the integrated scientific environment «Graph Model Workshop» is demonstrated.
The article describes the original software tools for an experimental estimation of computational complexity of software solutions for problems on graph models of systems. The classes of the solved problems and the tools for analysis of results are listed. The method based on selection of graph models by their structural complexity is introduced.