An asynchronous distributed discrete event simulation algorithm for cyclic circuits using a data-flow network

The authors propose a mathematically proven algorithm for asynchronous, distributed, discrete-event simulation of circuits containing feedback loops. Every component of the circuit is represented through an executable model and the flow of information is expressed through message communication. Simulation time is encapsulated in the model description and encoded in the messages. A data-flow network is synthesized based on the connectivity of the components in the circuit that computes a quantity time of next event for every component. This quantity permits the corresponding model to execute asynchronously as far ahead in simulation time as possible and yet guarantees correctness. The network ensures that any simulation process executing on a distributed processing environment that has sufficient information to simulate can execute while also avoiding deadlock. The algorithm has been verified through an implementation on loosely coupled parallel processor.<
>