Multi-Objective Multicast Environments for QoS Routing and a New Crossover with No Maximum Delay Constraint

G.M.B. Oliveira and S.S.B.V. Vita (Brazil)

Keywords

Multiobjective Optimization, Genetic Algorithms, NSGA, NSGAII, Multicast Routes, Quality of Service.

Abstract

Multicasting routing involves concurrently data transmission from a source to a subset of destination nodes in a computer network. Quality of Service (QoS) intends to guarantee that the end-to-end communication is obtained with the appropriate service level for each application. Several researchers have investigated genetic algorithms-based models for multicast route computation with QoS requirements. The evolutionary models proposed here use multi-objective approaches in a Pareto sense to solve this problem and to deal with the inheriting multiple metrics involved in QoS proposal. Basically, we construct two multicast QoS routing algorithms; the first one based on NSGA (Non-dominated Sorting Genetic Algorithms) and the second one based on NSGA-II. Three different pairs of objectives were evaluated: one is related to cost links and the other to delay links. We also investigated two different crossover processes. The first employs exactly the same strategy used in previous single-objective models, in which there is a need to pre specify a maximum delay constraint. The second gives the same probability to use either delay or cost metrics when a new multicast tree is generated and there is no need to specify a constraint. Results show that the best environment was obtained using NSGA-II and the equal delay/cost probability-based crossover.

Important Links:



Go Back