Space-time deterministic graph rewriting

Pablo Arrighi,Marin Costes,Gilles Dowek,Luidnel Maignan

2024-10-07

Abstract:We study non-terminating graph rewriting models, whose local rules are applied non-deterministically -- and yet enjoy a strong form of determinism, namely space-time determinism. Of course in the case of terminating computation it is well-known that the mess introduced by asynchronous rule applications may not matter to the end result, as confluence conspires to produce a unique normal form. In the context of non-terminating computation however, confluence is a very weak property, and (almost) synchronous rule applications is always preferred e.g. when it comes to simulating dynamical systems. Here we provide sufficient conditions so that asynchronous local rule applications conspire to produce well-determined events in the space-time unfolding of the graph, regardless of their application orders. Our first example is an asynchronous simulation of a dynamical system. Our second example features time dilation, in the spirit of general relativity.

Discrete Mathematics,Distributed, Parallel, and Cluster Computing,General Relativity and Quantum Cosmology,Dynamical Systems,Cellular Automata and Lattice Gases

What problem does this paper attempt to address?

This paper attempts to address the issue of achieving spatiotemporal determinism in non-terminating graph rewriting models. Specifically, the authors investigate a graph rewriting model where local rules can be applied non-deterministically but still exhibit strong determinism. For terminating computations, if the system satisfies confluence, the final result will be deterministic. However, in the context of distributed, non-terminating computations, confluence alone is insufficient to ensure consistency and determinism.

The authors propose two forms of spatiotemporal determinism:
1. **Weak Spatiotemporal Determinism**: Focuses on maintaining the consistency of certain events under all possible evaluation strategies.
2. **Strong Spatiotemporal Determinism**: Considers a broader category of events, which can be viewed as "partial results" of the former category of events.

Additionally, the authors demonstrate through several examples how asynchronous methods can be used to simulate dynamical systems and introduce dependency graphs (DAGs) to capture the relative progress between computational regions. Finally, they also show how such DAGs can be used to achieve effects similar to time dilation in general relativity.

Overall, this paper aims to address the issues of consistency and determinism in asynchronous computation processes and explore their potential applications in theoretical physics and efficient parallel computing.