Title: Mimicking Nature for Designing Robust Congestion Control Mechanisms in Self-Organized Autonomous Decentralized Networks Acronym: MiND2C Duration: 24 months Starting Date: December 1st, 2008 Funded by: Cyprus Research Promotion Foundation Budget: 99.960 EU
Nature, an everlasting source of inspiration, has been very successful in solving complex problems. The urge of living organisms toward survival often drives the emergence of intelligent behaviors which motivates the theory, design and application of Computational Intelligence (CI) paradigms e.g., Swarm Intelligence (SI). The mimicking of nature provides (alternative) solution techniques to problems that were not (satisfactorily) resolved by other traditional techniques. These new techniques are being applied successfully to a variety of scientific and engineering dynamic problems. We draw inspiration from nature for controlling the next generation communication networks, which are moving towards autonomous and decentralized infrastructures (e.g., ad-hoc, sensor networks) that are typically unstructured and tightly constrained in terms of power, computation, storage, and communication. The unpredictable nature of Autonomous Decentralized Networks (ADNs) experienced by topology modifications (due to node failures and mobility) along with traffic load variations and link capacity fluctuations (caused by sensing the environments and node interactions) may lead to congestion. Congestion causes energy waste, throughput reduction, increased delays and information loss leading to deterioration of the offered QoS and decrease of network lifetime. Research on ADNs has mainly focused on protocols and algorithms for applications in which network performance assurances are not considered essential, mainly due to the need to first solve the more rudimentary problems, as e.g. Medium Access Control and Routing. However, for many critical applications e.g., healthcare monitoring, performance assurances are crucial. The lack of solutions to promote dependability as well as to support QoS in ADNs necessitates robust and self-adaptive network control strategies. In this project, we will formulate the problem of congestion in ADNs and indentify QoS requirements for critical applications. We will investigate the usefulness of nature inspired techniques in controlling congestion in complex man-made systems. For example, using SI ideas for mimicking the behavioral tendencies of social insect colonies and other animal societies which attain global properties (robustness, self-adaptation), achieved collectively without explicitly programming them into individual nodes. We will design nature-inspired congestion control (CC) strategies for ADNs involving minimal exchange of information and computation burden which are simple to implement at the individual node. The constitution of multidisciplinary and experienced consortium indicates the presence of world-class expertise in all major segments of the work (communication networks, congestion control, Computational Intelligence, Swarm Intelligence) guarantees that the scientific and technological objectives of the project will be met. The consortium brings fresh ideas and contributes toward new research directions and the application of ADNs for critical applications. |
