There are several challenges associated with modeling smart factories, particularly in terms of the multi agent system. Hyperagent modeling, a technique involving fractal Petri nets, can significantly enhance the flexibility and sustainability of smart factories. By coordinating and managing multiple meta agents, hyperagents can optimize resource allocation, reduce waste, and improve efficiency. This approach can be applied to various scales of granularity, from individual machines to entire production lines. Smart factories are leveraging their sustainability through elastic computing-supported processes such as "folding" and "unfolding" of hyperagents modeled by fractal Petri nets. The paper is discussing the application of elastic computing and hyperagents in smart factories, specifically using a theoretical framework and a method of hyperagent release through folding of Petri nets.

The problem of automatic control of elastic computing granularity can be solved based on self-similar (fractal) network patterns. These patterns can control the level of granularity and ensure quick and efficient response to user demand, leading to optimal performance and minimized resource usage. For controlling and programming elastic computing granularity, the Fractal Abstract Machine (FAM) is proposed. It is also a theoretical framework for designing and modeling elastic computer systems, as well as for simulated elastic granularity, in order to solve optimization problems. Currently, FAM is investigating how to regulate an elastic network model while allocating containers and scheduling tasks in distribution networks. The correspondent example is considered. The idea of fractal computing, which controls elastic computing granularity, is explored.

The emerging paradigm shift for analysis of elastic systems is caused by demands of reorganization behavior of systems at run-time. For solution this problem a mathematical framework for graph-based dynamic analysis is introduced. The essential of the framework is a definition of the generalized graph-based model of the elastic system. Bond graphs, Fractal Petri nets, fractal graphs, and their matrix presentation are included in the framework and inherit technique of generalized model.

Pattern-type reachability analysis extends Petri nets analysis by constructing a reachability graph from the patterns which are the internal emergent properties of the modeling system. The process construction of the reachability tree takes into account Place and Transition invariants. Vertices of the pattern are markings satisfy a some set of constraints. The elastic model of the Dinning Philosophers based on Fractal Petri nets and its pattern-type analysis are considered. Fractal algebra operations implements by matrix operations. It makes possible to consider a state space explosion problem as patterns emergency and reachability.