Complexity of Networks

Russell K. Standish

DOI: https://doi.org/10.48550/arXiv.cs/0508075

2005-08-17

Abstract:Network or graph structures are ubiquitous in the study of complex systems. Often, we are interested in complexity trends of these system as it evolves under some dynamic. An example might be looking at the complexity of a food web as species enter an ecosystem via migration or speciation, and leave via extinction. In this paper, a complexity measure of networks is proposed based on the {\em complexity is information content} paradigm. To apply this paradigm to any object, one must fix two things: a representation language, in which strings of symbols from some alphabet describe, or stand for the objects being considered; and a means of determining when two such descriptions refer to the same object. With these two things set, the information content of an object can be computed in principle from the number of equivalent descriptions describing a particular object. I propose a simple representation language for undirected graphs that can be encoded as a bitstring, and equivalence is a topological equivalence. I also present an algorithm for computing the complexity of an arbitrary undirected network.

Information Theory

What problem does this paper attempt to address?

This paper attempts to solve the problem of complexity measurement in the study of complex systems with network or graph structures. Specifically, the author proposes a network complexity measurement method based on the information - content paradigm. To apply this paradigm, two things need to be determined: a representation language in which symbol strings from a certain alphabet are used to describe or represent the objects under consideration; and a method to determine whether two descriptions refer to the same object. By setting these two points, in principle, the information content of an object can be calculated from the number of equivalent descriptions of a specific object.

The representation language proposed in the paper is a simple implementation for undirected graphs, which can be encoded as bit - strings, and the equivalence is topological equivalence. In addition, the author also introduces an algorithm for calculating the complexity of any undirected network. The goal of the paper is to provide a theoretically satisfactory complexity measurement method, although this measurement depends on the choice of observer \(O\) and the representation language.

The paper further explores the relationship between the proposed complexity measurement and thermodynamic entropy, proposes a new compression complexity measurement method (called \(z\) - complexity), and evaluates the effectiveness and applicability of these measurement methods by analyzing different types of networks (such as random networks and regular networks). Ultimately, the author hopes that these measurement methods can be applied to networks generated by dynamic processes, especially networks in the evolutionary process, in order to better understand the evolution trends of complex systems.