Use of Learning, Game Theory and Optimization as Biomimetic Approaches for Self-Organization in Heterogeneous Networks

Abstract

In this chapter, we present the use of several bio inspired approaches called biomimetics for the design of Self-organization (SO) in heterogeneous network scenarios. Mainly these approaches are further categorized in indirect biomimetics and direct biomimetics depending on whether the inspiration from the biological systems is used indirectly or it is applied as it is to design SO. Under the umbrella of indirect biomimetics we discuss in detail the emerging paradigms in learning theory that have been recently shown to have strong potential for designing SO solution in heterogeneous networks. In the second part of the chapter, we investigate a rather under explored paradigm of direct biomimetic. Building on a case study of a self-organising systems in nature we extract generic SO design principles that can be used as a direct biomimetic approach for designing distributed, scalable and agile solutions to many problems in complex heterogamous networks. We demonstrate this direct biomimetic approach through a use case of a heterogeneous network scenario with outdoor fixed relays. By exploiting one to one mapping between a natural SO system and our system model we systematically apply the bio-inspired design principle directly and obtain a distributed SO solution to the problem under consideration. The performance of this solution is evaluated through numerical results and substantial gains are observed. Finally we conclude this chapter with remarks on some important considerations and limitations of the use of biomimetic approaches.