Abstract Problems of global optimization arise in many areas of modern science and have an ample number of applications in engineering, industry and economics as for circuit design, process planning, and scheduling. There are various heuristic methods to solve these problems such as bio-inspired approaches like Evolutionary Algorithms or Particle Swarm Optimization but also simpler approaches as Simulated Annealing or Threshold Accepting. In this article we introduce new analytical approaches to understand the underlying principles in particular of parallel evolutionary methods and their parameters better. More specific, we focus on the benefit of migration in the island model of evolutionary algorithms and on the influence of migration parameters. Zusammenfassung Probleme globaler Optimierung treten in vielen Gebieten der modernen Wissenschaften auf und haben unzählige Anwendungen in den Ingenieurwissenschaften, der Industrie und der Ökonomie wie beispielsweise für den Schaltkreisentwurf, zur Prozessplanung sowie zum Scheduling. Es gibt verschiedenste heuristische Methoden, um diese Probleme zu lösen, etwa bio-inspirierte Ansätze wie Evolutionäre Algorithmen oder Partikelschwarm-Optimierung oder einfachere Ansätze wie Simuliertes Ausglühen oder Schwellenakzeptanz. In diesem Artikel stellen wir neue Ansätze zum besseren Verständnis der tieferliegenden Prinzipien insbesondere von parallelen evolutionären Methoden sowie ihrer Parameter vor. Im Speziellen untersuchen wir die Nützlichkeit von Migration in parallelen genetischen Algorithmen im Inselmodell paralleler evolutionärer Algorithmen und den Einfluss von Migrationsparametern.
[1]
Günter Rudolph,et al.
Takeover time in parallel populations with migration
,
2006
.
[2]
Christian Blum,et al.
Metaheuristics in combinatorial optimization: Overview and conceptual comparison
,
2003,
CSUR.
[3]
P Bourgine,et al.
Towards a Practice of Autonomous Systems
,
1992
.
[4]
Gerhard W. Dueck,et al.
Threshold accepting: a general purpose optimization algorithm appearing superior to simulated anneal
,
1990
.
[5]
David E. Goldberg,et al.
On the Scalability of Parallel Genetic Algorithms
,
1999,
Evolutionary Computation.
[6]
James Kennedy,et al.
Particle swarm optimization
,
2002,
Proceedings of ICNN'95 - International Conference on Neural Networks.
[7]
Vijay V. Vazirani,et al.
Approximation Algorithms
,
2001,
Springer Berlin Heidelberg.
[8]
Marco Tomassini,et al.
Spatially Structured Evolutionary Algorithms: Artificial Evolution in Space and Time (Natural Computing Series)
,
2005
.
[9]
Dirk Sudholt,et al.
Experimental supplements to the theoretical analysis of migration in the Island model
,
2010,
PPSN 2010.
[10]
Dirk Sudholt,et al.
The benefit of migration in parallel evolutionary algorithms
,
2010,
GECCO '10.
[11]
C. D. Gelatt,et al.
Optimization by Simulated Annealing
,
1983,
Science.
[12]
Enrique Alba,et al.
Parallel evolutionary algorithms can achieve super-linear performance
,
2002,
Inf. Process. Lett..