Cognitive Approaches for Intelligent Networks

Cognitive Approaches for Intelligent Networks

T.R. Gopalakrishnan Nair
Copyright: © 2015 |Pages: 11
DOI: 10.4018/978-1-4666-5888-2.ch013
(Individual Chapters)
No Current Special Offers

Chapter Preview



Application of intelligence into networking is one of the ambitious targets in the research field, which is expected to produce a transformational impact on the performance of current networks. One of the major demands for future intelligent network is its ability to support the vastly increasing number of applications. Future smart networks also call for a successful scenario where it will co-exist with the present infrastructure. The existing infrastructure tends to lack the ability to learn and understand the ways and means of operation of applications, which often leads to various anomalies. In this context, it is necessary to view the network as a whole and establish appropriate intelligence in and around the nodes to recognize the properties of the environment of a communication point in relation to multiple hop circles around it. Currently network systems are much dependent upon old approaches of design having predetermined forward routing technology which produces usually non-optimal performance (Ciancetta, Colombo, Lavagnolo, & Grillo, 1999).

The routing decisions in the existing networks are generally based on table driven systems, provided at the node level. This system usually has no awareness of the environment around it. This is where intelligent networks can play a significant role. Many researchers are working in the area of autonomic network (AN) since 2003. The companies like Motorola and IBM were some of the players in this domain. Aspects of game theory, probability, linear programming, evolutionary algorithm, genetic algorithm, artificial immune system, artificial intelligence and many more stochastic approaches have been applied to achieve the awareness and learning capabilities in the network. Once there is awareness about the network, routing can be performed effectively. In order to meet the demand for improved network, the nodes need to be intelligent and capable of making decisions on their own. The current nodes and systems do not have the information regarding the topology which might have been formed during the routing. This will demand a process for collecting the knowledge about the environment. Hence the nodes need to have a mechanism to collect the data and become aware of the vital parameters of the network and if required, communicate this intelligently to other participating nodes. This will enable the network to realise the ability to learn, remember and reason out in a way as presented in Figure 1, as Cognitive cycle, which was initially proposed by Mitola (2001).

Figure 1.

Cognitive cycle


Key Terms in this Chapter

Information Base: The learnt information that is stored in a database for future references.

Genetic Algorithm: An evolutionary approach applied in systems based on gene property of human being.

Cognitive Network: A communication network involving an approach that produces awareness in networks through learning, planning, decision making and reasoning.

Reasoning: Another essential concept to be incorporated in any intelligent systems to derive the output / result based on learning.

Forward Channel: The links suitable for forwarding the data.

Intelligent Network: A network which learns the environment situation and adapts to changes based on user requirement.

Learning: An essential operation of acquiring, processing and storing information required by any intelligent system for evolution.

QoS: Quality of Service Parameters used to understand the network environment.

Complete Chapter List

Search this Book: