ZCS: A Zeroth Level Classifier System

@article{Wilson1994ZCSAZ,
  title={ZCS: A Zeroth Level Classifier System},
  author={Stewart W. Wilson},
  journal={Evolutionary Computation},
  year={1994},
  volume={2},
  pages={1-18}
}
A basic classifier system, ZCS, is presented that keeps much of Holland's original framework but simplifies it to increase understandability and performance. ZCS's relation to Q-learning is brought out, and their performances compared in environments of two difficulty levels. Extensions to ZCS are proposed for temporary memory, better action selection, more efficient use of the genetic algorithm, and more general classifier representation. 
View on MIT Press
eskimo.com
565 Citations
Highly Influential Citations
88
Background Citations
180
Methods Citations
182
Results Citations
12

565 Citations

Citation Type

Citation Type

A Simple Payoff-Based Learning Classifier System
TLDR
A simple payoff-based learn- ing classifier system with which to explore aspects of fitness sharing in such systems, a further aim being to achieve similar performance to accuracy-based learning classifier systems.
State of XCS Classifier System Research
TLDR
This paper reviews recent research on XCS with respect to representation, internal state, predictive modeling, noise, and underlying theory and technique.
ZCS Redux
TLDR
By using simple difference equation models of ZCS, it is shown that this system is capable of optimal performance subject to appropriate parameter settings, and this is demonstrated for both single- and multistep tasks.
Learning classifier systems: a survey
TLDR
The present contribution is to describe the state-of- the-art of LCSs, emphasizing recent developments, and focusing more on the sequential decision domain than on automatic classification.
Foundations of learning classifier systems: An introduction
TLDR
This work presents a kind of rule-based system with general mechanisms for processing rules in parallel, for adaptive generation of new rules, and for testing the effectiveness of existing rules.
A Self-Adaptive Classifier System
TLDR
It is shown that the mutation rate, learning rate, discount factor and tax rate can be developed along with an appropriate solution/rule-base, resulting in improved performance over results using fixed rate parameters.
Two Views of Classifier Systems
This work suggests two ways of looking at classifier systems; as Genetic Algorithm-based systems, and as Reinforcement Learning-based systems, and argues that the former is more suitable for…
A simple accuracy-based learning classifier system
TLDR
This paper presents a simple accuracy-based learning classifier system with which to explore aspects of accuracy- based fitness in general, and describes and modelled, before being implemented and tested on the multiplexer task.
A brief history of learning classifier systems: from CS-1 to XCS and its variants
  • L. Bull
  • Computer Science
    Evol. Intell.
  • 2015
TLDR
This paper gives an overview of the evolution of Learning Classifier Systems up to XCS, and of some of the subsequent developments of Wilson’s algorithm to different types of learning.
A Learning Classifier System with Mutual-Information-Based Fitness
TLDR
A new variety of learning classifier system (LCS), called MILCS, which utilizes mutual information as fitness feedback, and is specifically designed for supervised learning.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 49 REFERENCES
A Critical Review of Classifier Systems
TLDR
This work presents a kind of rule-based system with general mechanisms for processing rules in parallel, for adaptive generation of new rules, and for testing the effectiveness of existing rules.
Classifier Fitness Based on Accuracy
TLDR
A classifier system, XCS, is investigated, in which each classifier maintains a prediction of expected payoff, but the classifier's fitness is given by a measure of the prediction's accuracy, making it suitable for a wide range of reinforcement learning situations where generalization over states is desirable.
Genetic and Non-Genetic Operators in ALECSYS
  • M. Dorigo
  • Computer Science, Mathematics
    Evolutionary Computation
  • 1993
TLDR
This paper introduces the following original features of ALECSYS, a parallel version of a standard learning classifier system (CS), and presents simulation results of experiments run in a simulated two-dimensional world in which a simple agent learns to follow a light source.
Representing Attribute-Based Concepts in a Classifier System
TLDR
This paper describes some straightforward binary encodings for attribute-based instance spaces that give classifier systems the ability to represent ordinal and nominal attributes as expressively as most symbolic machine learning systems, without sacrificing the building blocks required by the genetic algorithm.
Hierarchical Credit Allocation in a Classifier System
TLDR
This work suggests an alternative form for the algorithm and the system's operating principles designed to induce behavioral hierarchies in which modularity of the hierarchy would keep all bucket-brigade chains short, thus more reinforceable and more rapidly learned, but overall action sequences could be long.
Default hierarchy formation and memory exploitation in learning classifier systems
TLDR
A class of problems that isolate memory exploitation from other aspects of LCS behavior is developed, showing that an LCS can form rule sets that exploit memory and does not form optimal rule sets because of a limitation in its allocation of credit scheme.
Intelligent behavior as an adaptation to the task environment ; Part II.
TLDR
This dissertation argues that examining more closely the way animate systems cope with real-world environments can provide valuable insights about the structural requirements for intelligent behavior.
Knowledge Growth in an Artificial Animal
TLDR
This paper describes work using an artificial, behaving, animal model (termed an “ani-mat”) to study intelligence at a primitive level, and wishes to provide the animat with adaptive mechanisms which yield rapid and solid improvement but themselves contain minimal a priori information.
An optimization-based categorization of reinforcement learning environments
TLDR
The paper discusses the special cases when either h = 0 or = 1 in detail, describes some theoretical bounds on h and re-explores a well-known reinforcement learning environment with this new notation.
The Fuzzy Classifier System: A Classifier System for Continuously Varying Variables
...
1
2
3
4
5
...