Graph models and operations in evaluating microservice solutions

Abstract

This article discusses how to optimize command-line interface (CLI) test coverage given complex specifications and limited resources. Traditional testing methods often fail to consider real-world scenarios, which can lead to inefficient use of resources to cover rarely used functionality. The authors review the current state of the problem and present a proposed set-theoretic model that takes into account the topological properties of the specification graph, as well as algorithmic and informational components. The set of test scenarios is generated using topological breadth-first and depth-first search (BFS and DFS) methods. A key feature of the model is the inclusion of a usage awareness parameter, which, based on dynamic call tracking, assigns awareness weights to graph nodes. The process of selecting the optimal test set is formalized using Shannon's information entropy, where the selection strategy at each iteration maximizes the removal of uncertainty regarding the system's operational state. The concept of average weighted coverage and a metric for relative awareness gain are introduced. The proposed model provides a transition from structural code coverage to user-value coverage. doi 10.54708/19926502_2026_30211249