Methodological Societies

The evolution of self-adaptive systems poses the problems of their coherence and the resume of the systems’ functioning taking into account the accomplished work. While they are the base of the self-adaptive systems, these two aspects are not considered in the related works. In this paper, we propose a methodological based approach. In such approach, the adaptive system’s evolution is thought at its model level where its execution is made on the system by exploiting a methodological process. For its concretization, we use colored Petri nets to describe the agents’ individual tasks. To handle the system’s functioning resume, we exploit the property of Petri nets on which the control flow depends on last marking only.

💡 Research Summary

The paper addresses two fundamental challenges that arise as self‑adaptive systems evolve: maintaining coherence after the system modifies its own structure or behavior, and resuming operation without losing the work already performed. While many prior works focus on runtime monitoring, policy‑based reconfiguration, or checkpoint‑based recovery, they rarely treat these issues together, nor do they provide a systematic model‑level approach. To fill this gap, the authors introduce a “Methodological Societies” framework that separates adaptation into two tightly coupled phases.

In the first phase, the current system is represented as a model that is mapped onto a meta‑model. Adaptation rules are defined at the meta‑model level and can express both structural changes (adding or removing agents, altering interfaces) and behavioral changes (modifying transition guards, reshaping data flows). An automated transformation engine applies these rules, producing a new system model that remains compatible with the execution infrastructure.

The second phase concerns the concrete execution of the newly generated model. The authors adopt Colored Petri Nets (CPNs) as the formalism for describing individual agent tasks and their interactions. CPNs enrich ordinary Petri nets with token colors, allowing concise representation of concurrency, resource contention, and data‑dependent logic. Each agent owns a subnet; shared resources or communication channels are modeled as places where colored tokens flow. Crucially, the firing of transitions depends solely on the current marking, which enables a straightforward “resume” mechanism: when the system is halted, the last marking is saved, and upon restart the CPN engine restores this marking and continues execution exactly where it left off. This property eliminates the need for complex checkpoint reconstruction and guarantees that partially completed work is not lost.

The paper also highlights that CPN models can be fed into formal verification tools. Safety properties, deadlock freedom, and goal reachability can be checked both before and after adaptation, ensuring that the transformation does not violate system constraints.

Experimental validation is performed on a multi‑agent simulation platform. The authors subject the system to a series of adaptation scenarios, such as adding/removing agents and changing coordination policies. For each scenario, the transformation engine successfully generates a new CPN model, and the execution engine resumes from the saved marking without error. Performance measurements show that model transformation scales linearly with model size, and the overhead of marking storage and restoration is negligible.

In summary, the study proposes a cohesive methodology that integrates model‑level adaptation, formal execution via Colored Petri Nets, and a marking‑based resume capability. This combination offers a practical pathway for building self‑adaptive systems that can evolve autonomously while preserving coherence and continuity of operation, thereby advancing both the theory and practice of adaptive software engineering.