Feasibility of Evolutionary Design for Multi-Access MAC Protocols

This paper explores the feasibility of evolving Medium Access Control (MAC) protocols using evolving finite state machines. Network protocols formulated as genotypes are evaluated through an evaluation framework and made to compete under the pressure of a desired performance fitness. Specifically, the evolution of two well-known protocols, Slotted-ALOHA and Carrier Sense Multiple Access (CSMA) is explored. Using a C based protocol fitness evaluator and a Java based evolution fabric, we show that a generalized state machine is able to evolve complex MAC behavior when time or channel sensing modalities are provided to the network nodes. We show that protocol synthesis takes place under a wide range of network parameters, loading conditions as well as limited/full evolutionary degrees of behavioral freedom. Performed adaptability experiments prove feasibility of the framework to adapt with runtime changes of single or multiple system parameters. The ultimate goal of this research is to develop an alternative design paradigm for protocols in heterogeneous network and traffic scenarios.