Modeling a holonic agent based solution by Petri nets

One of the key design issues for distributed systems is to find proper planning and coordination mechanisms when knowledge and decision capabilities are spread along the system. This contribution refers holonic manufacturing execution systems and highlights the way a proper modeling method – Petri nets – makes evident certain problems that can appear when agents have to simultaneously treat more goals. According to holonic organization the planning phase is mainly dependent on finding an appropriate resource allocation mechanism. The type of weakness is established by means of the proposed Petri net models and further proved by simulation experiments. A solution to make the holonic scheme avoid a failure in resource allocation is mentioned, too.