A Categorical Approach to Threshold Agent Networks

Threshold agent networks (TANs) constitute a discretized modification of threshold (also known as neural) networks that are appropriate for modeling computer simulations. In this paper morphisms are introduced between TANs, thus forming a category TAN. Several of the properties of this category are explored. A generalization of TANs to allow more flexible interactions between the agents is then introduced. The new objects are also endowed with appropriate morphisms and the properties of the newly obtained category are compared with the properties of its older subcategory. Mathematics Subject Classifications (2000): Primary: 93A05, 18B20; secondary: 92B20, 94Cxx, 18A35. Key words: finite dynamical systems, automata networks, threshold networks, neural networks, threshold agent networks, categories of networks, products, equalizers, limits.

• Type: Journal
• Year: 2004
• Where: ACS
• Authors: George Voutsadakis
• Keywords: ACS 2004, ACS 2007, Category Tan, TANs, Threshold Agent