Font Size: a A A

Research On Multicast Routing Problem Based On Multi-objective Optimization

Posted on:2019-06-10Degree:MasterType:Thesis
Country:ChinaCandidate:W M QuanFull Text:PDF
GTID:2348330569488931Subject:Electronics and Communications Engineering
Abstract/Summary:PDF Full Text Request
Since the birth of the Internet,the scale of its users and market demand has been growing rapidly.The world is now in a new era of information explosion.In people's daily life,various types of multimedia applications can be found everywhere.Multimedia services with high-traffic and high-bandwidth enrich people's lives and also have a huge impact on the network.The emergence of multicast technology precisely meets the needs of such network services.Multicast routing is one of the key scientific issues for implementing multicast,and is also an important research subject in the current structure of IP network and future network.The contradiction and correlation between quality of service parameters in multicast routing problem satisfy the characteristics of multi-objective optimization problems.This paper studies the multicast routing problem based on multi-objective optimization and proposes two kinds of improved algorithms to solve multicast routing problem with the average end-to-end delay and packet loss rate as the optimization objectives.The main contents are as follows:1)Based on the multi-objective evolutionary algorithm framework with dominance relations,this paper proposes a multi-objective evolutionary algorithm based on pheromone composition.The algorithm combines the advantages of ant colony optimization and Jaya,uses the pheromone-directed construction map strategy to accelerate the convergence of the algorithm.By approaching the optimal solution and distancing the worst solution,a multi-objective learning strategy based on the Jaya algorithm is proposed,and the Lévy random perturbation operator is introduced to increase the diversity of the solution.Through experimental simulation and comparison,this algorithm effectively solves the problem of multi-objective multicast routing with minimum average delay and packet loss rate under the bandwidth constraint.It is obviously better than NSGA-II algorithm and ENS-NDT algorithm in term of solving this problem.2)Based on the multi-objective evolutionary algorithm framework of the decomposition model,this paper proposes a MOEA/D algorithm that enhances the neighborhood search and elite inverse learning.This algorithm abandons the strategy of generating new solutions to subproblems in the original MOEA/D algorithm,and adopts an enhanced neighborhood search strategy,which improves the convergence of the algorithm.Besides,it introduces the idea of reverse learning into the multi-objective evolutionary algorithms and uses the reverse learning ability of elite individuals to increase the diversity of solutions.Through experimental simulation and comparison,it can be seen that the two algorithms proposed in this paper are superior to the existing four multi-objective evolutionary algorithms in solving multi-objective multicast routing problems with minimum average delay and packet loss rate under bandwidth constraints.Provide decision makers with richer and better quality alternative solutions.3)Based on the two multi-objective multicast routing algorithms proposed in this paper,a Software-Defined multi-objective multicast routing scheme is proposed,and a Software-Defined Networking multicast service system for real-time video streaming is designed and implemented.It can be seen from the experimental simulation and comparison that the Software-Defined multi-objective multicast routing strategy proposed in this paper can effectively solve the multicast routing problem in Software-Defined Networking and provide more efficient multicast services.
Keywords/Search Tags:Multicast Routing Problem, Combinatorial Optimization, Multi-objective Optimization, Evolutionary Algorithms, Software-Defined Networking
PDF Full Text Request
Related items