In this research paper we present an immunological algorithm (IA) to solve global numerical optimization problems for high-dimensional instances. Such optimization problems are a crucial component for many real-world applications. We designed two versions of the IA: the first based on binary-code representation and the second based on real values, called opt-IMMALG01 and opt-IMMALG, respectively. A large set of experiments is presented to evaluate the effectiveness of the two proposed versions of IA. Both opt-IMMALG01 and opt-IMMALG were extensively compared against several nature inspired methodologies including a set of Differential Evolution algorithms whose performance is known to be superior to many other bio-inspired and deterministic algorithms on the same test bed. Also hybrid and deterministic global search algorithms (e.g., DIRECT, LeGO, PSwarm) are compared with both IA versions, for a total 39 optimization algorithms.The results suggest that the proposed immunological algorithm is effective, in terms of accuracy, and capable of solving large-scale instances for well-known benchmarks. Experimental results also indicate that both IA versions are comparable, and often outperform, the state-of-the-art optimization algorithms.
Clonal Selection - An Immunological Algorithm for Global Optimization over Continuous Spaces
PAVONE, MARIO FRANCESCO;NICOSIA, GIUSEPPE
2012-01-01
Abstract
In this research paper we present an immunological algorithm (IA) to solve global numerical optimization problems for high-dimensional instances. Such optimization problems are a crucial component for many real-world applications. We designed two versions of the IA: the first based on binary-code representation and the second based on real values, called opt-IMMALG01 and opt-IMMALG, respectively. A large set of experiments is presented to evaluate the effectiveness of the two proposed versions of IA. Both opt-IMMALG01 and opt-IMMALG were extensively compared against several nature inspired methodologies including a set of Differential Evolution algorithms whose performance is known to be superior to many other bio-inspired and deterministic algorithms on the same test bed. Also hybrid and deterministic global search algorithms (e.g., DIRECT, LeGO, PSwarm) are compared with both IA versions, for a total 39 optimization algorithms.The results suggest that the proposed immunological algorithm is effective, in terms of accuracy, and capable of solving large-scale instances for well-known benchmarks. Experimental results also indicate that both IA versions are comparable, and often outperform, the state-of-the-art optimization algorithms.File | Dimensione | Formato | |
---|---|---|---|
pavone-jogo-2011.pdf
solo gestori archivio
Tipologia:
Documento in Post-print
Licenza:
Non specificato
Dimensione
1.09 MB
Formato
Adobe PDF
|
1.09 MB | Adobe PDF | Visualizza/Apri |
Nicosia-J-Global-Optimization-2012.pdf
solo gestori archivio
Tipologia:
Versione Editoriale (PDF)
Dimensione
1.1 MB
Formato
Adobe PDF
|
1.1 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.