Inproceedings836

Time-Scattered Heuristic for the Hardware Implementation of Population-based ACO

Bernd Scheuermann, Michael Guntsch, Martin Middendorf, Hartmut Schmeck

Published: 2004 September
Herausgeber: M. Dorigo et al.
Buchtitel: Ant colony Optimization and Swarm Intelligence. Proc. of the ANTS 2004 conference
Ausgabe: 3172
Reihe: LNCS
Seiten: 250-261
Verlag: Springer-Verlag

Referierte Veröffentlichung

BibTeX

Kurzfassung
We present a new kind of heuristic guidance as an extension to the Population-based Ant Colony Optimization (P-ACO) implemented in hardware on a Field Programmable Gate Array (FPGA). The heuristic information is obtained by transforming standard heuristic information into small time-scattered heuristic-vectors of favourable ant decisions. This approach is suited for heuristics which allow for an a priori calculation of the heuristics information. Using the proposed method, an ant can build-up a solution in quasi-linear time. Experimental studies measure the performance of the time-scattered heuristic. A comparison with the standard heuristic and candidate lists is also given.

Projekt

OptRek

Forschungsgruppe

Effiziente Algorithmen

Forschungsgebiet

Ameisenalgorithmen, Rekonfigurierbarkeit, Rechnerarchitektur, Rekonfigurierbare Gitter