Autonomously Evolving Communication Protocols

Abstract

While traditional telecommunication still relies on rigid, highly regulated and highly controlled communication protocols, with the emergence of new forms of networks (mobile ad hoc and delaytolerant networks, lacking central infrastructure and strict regulations) bio-inspired communication protocols have also found their way to success. In this chapter we introduce a nontraditional way of creating and shaping communication protocols, through an autonomous machine intelligence model, built upon on-line evolutionary methods such as natural selection and genetic programming. Creating a genetic programming language and a selection mechanism for multi-hop broadcast protocols in ad hoc networks, we show that this kind of approach can outperform traditional ones under given circumstances, offering a powerful alternative in the future.