The present study analyzes the software implementation of the concept of generating Turing machines that solve NP-hard problems. The paper describes a mathematical formulation of machine generation, presented as a solution of resource planning problems to be adapted to software for designing class schedules or, more generally, for network scheduling. The proposed software solutions for data representation are partially based on technical solutions for designing a class schedule. Basic strategies (procedural semantics of methods) for reducing searces (deep return, look ahead, etc.) utilize technical solutions for designing class schedules and other strategies (e.g., application of oracle) that have not been used for this purpose. The study is aimed at analyzing the intermediate results of software implementation of the concept of generating Turing machines that solve NP-hard problems. The paper proposes further plans for developing a software package for generating Turing machines into an open platform for teaching students the theory of algorithms and information technology, conducting experiments to determine the non-polynomial complexity of solving certain NP-hard problems generated by Turing machines.