Three architectures for continuous action

Authors:
Stewart W. Wilson
Affiliations:
Prediction Dynamics, Concord, MA and Department of General Engineering, The University of Illinois at Urbana-Champaign IL
Venue:
IWLCS'03-05 Proceedings of the 2003-2005 international conference on Learning classifier systems
Year:
2007

Citing 7
Cited 10

Introduction to Reinforcement Learning

Introduction to Reinforcement Learning
Extending XCSF beyond linear approximation

GECCO '05 Proceedings of the 7th annual conference on Genetic and evolutionary computation
Kernel-based, ellipsoidal conditions in the real-valued XCS classifier system

GECCO '05 Proceedings of the 7th annual conference on Genetic and evolutionary computation
Toward Optimal Classifier System Performance in Non-Markov Environments

Evolutionary Computation
Classifier fitness based on accuracy

Evolutionary Computation
An evolutionary function approximation approach to compute prediction in XCSF

ECML'05 Proceedings of the 16th European conference on Machine Learning
Toward a theory of generalization and learning in XCS

IEEE Transactions on Evolutionary Computation

On the effects of node duplication and connection-oriented constructivism in neural XCSF

Proceedings of the 10th annual conference companion on Genetic and evolutionary computation
Self-adaptive constructivism in Neural XCS and XCSF

Proceedings of the 10th annual conference on Genetic and evolutionary computation
A new approach to fuzzy LCSs in two-dimensional continuous multistep environment with continuous vector actions

Proceedings of the 10th annual conference on Genetic and evolutionary computation
QFCS: A Fuzzy LCS in Continuous Multi-step Environments with Continuous Vector Actions

Proceedings of the 10th international conference on Parallel Problem Solving from Nature: PPSN X
Classifier Conditions Using Gene Expression Programming

Learning Classifier Systems
Towards continuous actions in continuous space and time using self-adaptive constructivism in neural XCSF

Proceedings of the 11th Annual conference on Genetic and evolutionary computation
uQFCS: QFCS with unfixed fuzzy sets in continuous multi-step environments with continuous vector actions

Proceedings of the 11th Annual conference on Genetic and evolutionary computation
Fuzzy dynamical genetic programming in XCSF

Proceedings of the 13th annual conference companion on Genetic and evolutionary computation
XCSR with computed continuous action

AI'12 Proceedings of the 25th Australasian joint conference on Advances in Artificial Intelligence
Dynamical genetic programming in xcsf

Evolutionary Computation

Quantified Score

Hi-index	0.00

Visualization

Abstract

Three classifier system architectures are introduced that permit the systems to have continuous (non-discrete) actions. One is based on interpolation, the second on an actor-critic paradigm, and the third on treating the action as a continuous variable homogeneous with the input. While the last architecture appears most interesting and promising, all three offer potential directions toward continuous action, a goal that classifier systems have hardly addressed.