Assessment of and Algorithm for the Clustering Quality of Technological Solutions Using the Hamming Distance and Their Structure

In modern conditions, when designing and optimizing complex technical systems in the field of mechanical engineering, developers analyze a variety of alternative technical and technological solutions synthesized within the framework of the technical specifications. The number of potential technological solutions may include thousands or more potential options that require clustering (grouping) for subsequent analysis and selection; accordingly, clustering, as a way of reducing the number of options under consideration, is a key component for solving this problem. However, the efficiency of clustering directly depends on the chosen metric between objects and the method for assessing the quality of the resulting clusters. In the context of technical systems, where decisions are often encoded as binary or discrete vectors (presence or absence of a component, choice of node type, etc.), such classical metrics as Euclidean distance may not be informative enough. The article discusses the application of the inverse Hamming distance to assess the quality of clustering of alternative technological solutions, which allows moving from formal clustering to a structural-content model.