A novel strategy for the logic synthesis of asynchronous control circuits is presented. It is based on the structural theory of Petri nets and integer linear programming. Techniques that are capable of checking implementability conditions, such as complete state coding, and deriving a gate netlist to implement the specified behavior are presented. These techniques can handle Petri net specifications consisting of several thousands of transitions and provide a significant speed-up compared with techniques that have previously been proposed

A new method for the deadlock prevention problem in concurrent systems where a set of processes share a set of common resources in a conservative way is proposed. It can be applied to flexible manufacturing systems, modeled with Petri nets. In this paper, we present a set of important results related to the deadlock prevention problem in S4PR nets. First, a liveness characterization is introduced, establishing how deadlocks can be studied in terms of circular waits. Second, we show how a circular wait situation corresponds to a particular marking related to a siphon of the Petri net model. Finally, this last characterization is used to obtain an iterative method that successively forbids deadlock related states, synthesizing the control necessary to ensure a final live behavior. The method can be implemented by means of the solutions of a set of integer linear programming problems.

Banker's-like approaches to deadlock avoidance are based on a decision procedure to grant active processes resources using information about the maximum needs of resources that a process can request in order to ensure termination. The paper presents an extension of the classical Banker's algorithm to a class of flexible manufacturing systems modeled by means of Petri nets. These systems have two interesting characteristics from the application point of view. First, flexible routing of parts is allowed, and second, a multiset of resources is allowed to be used at each processing step. The decision procedure introduced is polynomial in the Petri net model size.