PT Unknown
AU Reichelt, D
TI Kommunikationsnetzwerkplanung unter Kosten- und Zuverlässigkeitsgesichtspunkten mit Hilfe von evolutionären Algorithmen
PD April
PY 2006
WP https://www.db-thueringen.de/receive/dbt_mods_00005635
LA de
DE Genetische Algorithmen; Reparaturheuristik; Netzwerkplanung; Evolutionäre Algorithmen; All-Terminal Reliability
AB Within the last 10 years business-to-business telecommunication gets a key 
success factor for many companies. The reliability of network topologies becomes 
a major design issue for modern communication network topologies. Furthermore, a 
network designer is faced by economic aspects. This thesis investigates the 
design of economic and reliable communication network topologies. Due to 
complexity of the design problem decision support systems are appropriated tools 
to support the network design process. This thesis proposes methods for decision 
support systems for the design of reliable and economic network topologies.
This thesis investigates furthermore the combinatorial network design problem 
with the objectives reliability and network setup cost. The reliability of a 
network topology is evaluated by the all-terminal reliability. Several methods 
for calculating the all-terminal reliability by an empirical comparison are 
examined.
Design problems with identical link reliabilities and design problems with 
different link options are investigated further on. The metaheuristic Genetic 
Algorithms (GA) is proposed to solve the combinatorial network design problem. 
First, GAs for single objective problems with constraints are proposed. 
Furthermore, multiobjective approaches optimizing both objectives in parallel 
are presented. An experimental comparison investigates the new GAs for common 
test problems and compares the results with existing approaches from literature.
ER