An artificial neural network representation for artificial organisms

An artificial neural network representation for artificial organisms

We introduce an artificial neural network (ANN) representation that supports the evolution of complex behaviors in artificial organisms. The strength and location of each connection in the network is specified by a connection descriptor. The connection descriptors are mapped directly into a bit-stri...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Collins, Robert J., Jefferson, David R.
Format: Tagungsbericht
Sprache:eng
Schlagworte:
Artificial Life
Connectivity Pattern
Genetic Algorithm
Hide Unit
Recurrent Network
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We introduce an artificial neural network (ANN) representation that supports the evolution of complex behaviors in artificial organisms. The strength and location of each connection in the network is specified by a connection descriptor. The connection descriptors are mapped directly into a bit-string to which a genetic algorithm is applied. We empirically compare this representation to other ANN-based representations in the complex AntFarm task.
ISSN:0302-9743
1611-3349
DOI:10.1007/BFb0029761