Investigation of Performance of Distributed Complex Systems Using Information-theoretic Means and Genetic Algorithms

Abstract

An investigation is conducted into performance measures to evaluate network-centric systems via their information or other flow properties. To approach this problem, concepts are borrowed from Graph Theory Information Theory, and current methods to analyze network-centric systems. A number of tools are presented to help better understand how to measure the flow in distributed networks. The efficacy of the proposed method is demonstrated by taking a known distributed paradigm (logistics system) and examining situations that produce maximum and minimum flow conditions. The optimization problem involving flow variables is computationally complex (NP-hard) and thus is determined via genetic algorithms.