A Pareto based approach with elitist learning strategy for MPLS/GMPS networks

Abstract

Abstract-Modern telecommunication networks are based on diverse applications that highlighted the status of efficient use of network resources and performance optimization. Various methodologies are developed to address multi-objectives optimization within the traffic engineering of MPLS/ GMPLS networks. However, Pareto based approach can be used to achieve the optimization of multiple conflicting objective functions concurrently. We considered two objective functions such as routing and load balancing costs functions. In the paper, we introduce a heuristics algorithm for solving multi-objective multiple constrained optimization (MCOP) in MPLS/ GMPLS networks. The paper proposes the application of a Pareto based particle swarm optimization (PPSO) for such network's type and through a comparative analysis tests its efficiency against another modified version; Pareto based particle swarm optimization with elitist learning strategy (PPSO_ELS). The simulation results showed that the former proposed approach not only solved the MCOP problem but also provide effective solution for exploration problem attached with PPSO algorithm.

An Improved Particle Swarm Algorithm for Multi-Objectives Based Optimization in MPLS/GMPLS Networks

“…The Link utilization l u is a number of traffic flows over the link. While the link capacity l c can be defined as a capacity of a link to handle unit traffic/number of traffic flows [2], [30]. By having these two functions, traffic load balancing (ϕ l ) cost can be illustrated as the link utilization according to its capacity as follow;…”

“…Selecting the optimal paths in multi-constraints networks is still a major challenge. Finding the best path from the list of feasible paths under real networks scenarios is called a Multi-Constrained Optimal Path (MCOP) problem [1], [2]. In actual networks, the optimal paths can be derived by computing the multiple objectives with…”

“…The QoS routing in such MPLS/ GMPLS networks is an MCOP optimization problem and is recognized as nondeterministic polynomial-time (NP) hard problem [2], [5]. Recently, numerous meta-heuristic optimization approaches (that include swarm, bio-inspired and genetic-based algorithms) have gained momentous popularity in various applications optimization, including the networks optimization.…”

A Modified Whale Optimization Algorithm for Enhancing the Features Selection Process in Machine Learning

“…Population-based techniques, such as simulated annealing, random generation, and metaheuristic algorithms, are examples of randomised algorithms. [5][6][7][8][9] Machine learning methods on the other hand are seen to solve the feature selection problem but with a great cost in computation, high complexities and premature convergence. In order to contain these drawbacks, metaheuristic algorithms were proposed, as they are good at dealing with these type of conditions.…”

Analysis of Artificial Intelligence-Based Metaheuristic Algorithm for MPLS Network Optimization

“…Routing protocols plays a vital role for the computation of optimal paths in MPLS-TE entity, where various algorithms can be used for optimal path (s) computation, dependent on objective functions. To compute optimal paths dependent on the objective functions leads to the concept of MPLS network optimization and is considered as NP hard [1], [2]. Recently, Artificial intelligence application-based optimization tools have got momentous appreciation for solving optimization problems in different fields.…”