Abstract

Multiprotocol label switched (MPLS) networks were introduced to enhance the network`s service provisioning and optimize its performance using multiple protocols along with label switched based networking technique. With the addition of traffic engineering entity in MPLS domain, there is a massive increase in the networks resource management capability with better quality of services (QoS) provisioning for end users. Routing protocols play an important role in MPLS networks for network traffic management, which uses exact and approximate algorithms. There are number of artificial intelligence-based optimization algorithms which can be used for the optimization of traffic engineering in MPLS networks. The paper presents an optimization model for MPLS networks and proposed dolphin-echolocation algorithm (DEA) for optimal path computation. For Network with different nodes, both algorithms performance has been investigated to study their convergence towards the production of optimal solutions. Furthermore, the DEA algorithm will be compared with the bat algorithm to examine their performance in MPLS network optimization. Various parameters such as mean, minimum /optimal fitness function values and standard deviation. Document type: Conference object

Abstract

Introduction of modern and diverse applications in telecommunication field has raised challenges in networking area regarding efficient use of network resources and with optimizing performance. Therefore MPLS/GMPLS (Generalized multiprotocol label switching) networks were introduced to provide a better quality of service to meet users' requirements as well as to optimize network resources. GMPLS networks use traffic engineering techniques for more efficient communication within the network and help to optimize network resources. This paper proposes BAT inspired metaheuristic algorithm for selecting an efficient route in MPLS/ GMPLS networks. In our investigation we considered routing costs as an objective function with goal to minimize it. The paper uses BAT algorithm with various levels of loudness parameter. The simulation results show performance improvements in MPLS/GMPLS networks of different size. Document type: Conference object