Exploiting distributed discrete-event simulation techniques for parallel execution of cellular automata

Abstract

The Cellular Automata (CA) paradigm is well-suited to model complex systems based on local rules of evolution such as those related to fluid-dynamics, crowd simulation, fire propagation and many more. In addition, CA can be profitably exploited as a support for different kinds of numerical approaches, such as finite element and finite volume methods. As the size of the problem increases, a cellular automaton can be easily parallelized through domain partitioning in order to scale up its execution. However, the performance and scalability of cellular automata executed on parallel/distributed machines are limited by the necessity of synchronizing all the nodes at each time step, i.e., a node can execute a new step only after all the other nodes have executed the previous one. This paper presents a preliminary study on how techniques taken from the Discrete-Event Simulation field can be adopted for the parallelization of CA on distributed memory architectures with the goal of reducing the synchronization burden. In particular, we combine the active/inactive cells technique, which is well-known in the CA context, with the concept of lookahead which, instead, is adopted in the field of distributed discrete-event simulation research.