The predecessor-existence problem for k-reversible processes

TL;DR

This paper investigates the computational complexity of the predecessor-existence problem in k-reversible processes on graphs, showing polynomial solutions for k=1 and special cases, but NP-completeness for k>1 in general.

Contribution

It proves NP-completeness of the problem for k>1, and provides efficient algorithms for trees and bounded-degree graphs, advancing understanding of these dynamical systems.

Findings

Predecessor Existence is polynomial for k=1.

NP-completeness established for k>1.

Efficient algorithms for trees and degree-bounded graphs.

Abstract

For k>=1, we consider the graph dynamical system known as a k-reversible process. In such process, each vertex in the graph has one of two possible states at each discrete time. Each vertex changes its state between the present time and the next if and only if it currently has at least k neighbors in a state different than its own. Given a k-reversible process and a configuration of states assigned to the vertices, the Predecessor Existence problem consists of determining whether this configuration can be generated by the process from another configuration within exactly one time step. We can also extend the problem by asking for the number of configurations from which a given configuration is reachable within one time step. Predecessor Existence can be solved in polynomial time for k=1, but for k>1 we show that it is NP-complete. When the graph in question is a tree we show how to solve it in O(n) time and how to count the number of predecessor configurations in O(n^2) time. We also solve Predecessor Existence efficiently for the specific case of 2-reversible processes when the maximum degree of a vertex in the graph is no greater than 3. For this case we present an algorithm that runs in O(n) time.