Reliable Communication Network Design with EvolutionaryAlgorithms

For the reliable communication network design (RCND) problem links areunreliable and for each link several options are available with dierent reliabilitiesand costs. The goal is to nd a cost-minimal communicationnetwork design that satises a pre-dened overall reliability constraint.This paper presents two new EA approaches, LaBORNet and BaBORNet,for the RCND problem. LaBORNet uses an encoding that representsthe network topology as well as the used link options and repairs infeasiblesolutions using an additional repair heuristic (CURE). BaBORNetencodes only the network topology and determines the link options byusing the repair heuristic CURE as a local search method.[...]

MoreLess

Year of publication:	2004-11-16
Authors:	Reichelt, Dirk ; Rothlauf, Franz
Institutions:	Betriebswirtschaftliches Institut <Mannheim> ; Universität <Mainz> / Lehrstuhl für Wirtschaftsinformatik und BWL
Subject:	Netzwerk \| Design \| design \| Evolutionärer Algorithmus

1 Introduction
2 Design of Cost-minimal Communication Networksunder Reliability Constraints
2.1 The Reliable Communication Network Design Problem
2.2 Reliability Measurements for Communication Networks
2.3 Existing Solution Approaches for the RCND problem
3 Evolutionary Algorithms using Repair Conceptsfor the Design of Reliable Communication Networks
3.1 Using Repair Heuristics for Evolutionary Algorithms
3.2 LaBORNet - A Lamarckian Based Optimizer for ReliableNetwork Design Problems
3.3 BaBORNet - A Baldwin Based Optimizer for Reliable Network Design Problems
3.4 CURE - A Deterministic Cut-Based Repair Heuristic
4 Experiments and Results
4.1 Test Problems
4.2 Experimental Design
4.3 Results
5 Summary and Conclusions
References

Persistent link: https://ebvufind01.dmz1.zbw.eu/10005868148