Performance and Efficiency of Memetic Pittsburgh Learning Classifier Systems

2009 ◽  
Vol 17 (3) ◽  
pp. 307-342 ◽  
Author(s):  
Jaume Bacardit ◽  
Natalio Krasnogor
Keyword(s):  
Learning Classifier Systems ◽  
Classification Rules ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
Classifier System ◽  
Learning Classifier ◽  
Rule Sets ◽  
Rule Set ◽  
Computational Resources ◽  
Robust To Noise

In this paper we empirically evaluate several local search (LS) mechanisms that heuristically edit classification rules and rule sets to improve their performance. Two kinds of operators are studied, (1) rule-wise operators, which edit individual rules, and (2) a rule set-wise operator, which takes the rules from N parents (N ≥ 2) to generate a new offspring, selecting the minimum subset of candidate rules that obtains maximum training accuracy. Moreover, various ways of integrating these operators within the evolutionary cycle of learning classifier systems are studied. The combinations of LS operators and policies are integrated in a Pittsburgh approach framework that we call MPLCS for memetic Pittsburgh learning classifier system. MPLCS is systematically evaluated using various metrics. Several datasets were employed with the objective of identifying which combination of operators and policies scale well, are robust to noise, generate compact solutions, and use the least amount of computational resources to solve the problems.

scholarly journals Smart Contact Tracing and Classifier System for Covid-19 Cases

2021 ◽  
Vol 9 (9) ◽  
pp. 1239-1245
Keyword(s):  
Operating System ◽  
Short Range ◽  
Contact Tracing ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
Classifier System ◽  
Incremental Method ◽  
Learning Classifier ◽  
Rule Set ◽  
Tracing Method

The growing shreds of evidence and spread of COVID-19 in recent times have shown that to effortlessly and optimally tackle the rate at which COVID-19 infected individuals affect uninfected individuals has become a pressing challenge. This demands the need for a smart contact tracing method for COVID-19 contact tracing. This paper reviewed and analysed the available contact tracing models, contact tracing applications used by 36 countries, and their underlined classifier systems and techniques being used for COVID-19 contact tracing, machine learning classifier methods and ways in which these classifiers are evaluated. The incremental method was adopted because it results in a step-by-step rule set that continually changes. Three categories of learning classifier systems were also studied and recommended the Smartphone Mobile Bluetooth (BLE) and Michigan learning classifier system because it offers a short-range communication that is available regardless of the operating system and classifies based on set rules quickly and faster.

ZCS Redux

2002 ◽  
Vol 10 (2) ◽  
pp. 185-205 ◽  
Author(s):  
Larry Bull ◽  
Jacob Hurst
Keyword(s):  
Genetic Algorithms ◽  
Difference Equation ◽  
Optimal Performance ◽  
Learning Classifier Systems ◽  
Rule Discovery ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
Classifier System ◽  
Learning Classifier ◽  
Simple Difference

Learning classifier systems traditionally use genetic algorithms to facilitate rule discovery, where rule fitness is payoff based. Current research has shifted to the use of accuracy-based fitness. This paper re-examines the use of a particular payoff-based learning classifier system—ZCS. By using simple difference equation models of ZCS, we show that this system is capable of optimal performance subject to appropriate parameter settings. This is demonstrated for both single- and multistep tasks. Optimal performance of ZCS in well-known, multistep maze tasks is then presented to support the findings from the models.

Bounding the Effect of Noise in Multiobjective Learning Classifier Systems

2003 ◽  
Vol 11 (3) ◽  
pp. 279-298 ◽  
Author(s):  
Xavier Llorà ◽  
David E. Goldberg
Keyword(s):  
Theoretical Model ◽  
Noisy Data ◽  
Learning Classifier Systems ◽  
Data Sets ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
Classifier System ◽  
Learning Classifier ◽  
Noisy Domains ◽  
The Impact

This paper analyzes the impact of using noisy data sets in Pittsburgh-style learning classifier systems. This study was done using a particular kind of learning classifier system based on multiobjective selection. Our goal was to characterize the behavior of this kind of algorithms when dealing with noisy domains. For this reason, we developed a theoretical model for predicting theminimal achievable error in noisy domains. Combining this theoretical model for crisp learners with graphical representations of the evolved hypotheses through multiobjective techniques, we are able to bound the behavior of a learning classifier system. This kind of modeling lets us identify relevant characteristics of the evolved hypotheses, such as overfitting conditions that lead to hypotheses that poorly generalize the concept to be learned.

Is a Learning Classifier System a Type of Neural Network?

1994 ◽  
Vol 2 (1) ◽  
pp. 19-36 ◽  
Author(s):  
Robert E. Smith ◽  
H. Brown Cribbs
Keyword(s):  
The Other ◽  
Adaptive Genetic Algorithm ◽  
Diverse Population ◽  
Distinguishing Characteristic ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
Classifier System ◽  
System A ◽  
Learning Classifier ◽  
The Relationship

This paper suggests a simple analogy between learning classifier systems (LCSs) and neural networks (NNs). By clarifying the relationship between LCSs and NNs, the paper indicates how techniques from one can be utilized in the other. The paper points out that the primary distinguishing characteristic of the LCS is its use of a co-adaptive genetic algorithm (GA), where the end product of evolution is a diverse population of individuals that cooperate to perform useful computation. This stands in contrast to typical GA/NN schemes, where a population of networks is employed to evolve a single, optimized network. To fully illustrate the LCS/NN analogy used in this paper, an LCS-like NN is implemented and tested. The test is constructed to run parallel to a similar GA/NN study that did not employ a co-adaptive GA. The test illustrates the LCS/NN analogy and suggests an interesting new method for applying GAs in NNs. Final comments discuss extensions of this work and suggest how LCS and NN studies can further benefit each other.

USING ACCURACY-BASED LEARNING CLASSIFIER SYSTEMS FOR ADAPTABLE STRATEGY GENERATION IN GAMES AND INTERACTIVE VIRTUAL SIMULATIONS

2009 ◽  
Vol 10 (04) ◽  
pp. 365-390
Author(s):  
ALEJANDRO JUAN ◽  
RICHARD W. PAZZI ◽  
AZZEDINE BOUKERCHE
Keyword(s):  
Machine Learning ◽  
Genetic Algorithm ◽  
Training Session ◽  
Learning Classifier Systems ◽  
Machine Learning Techniques ◽  
Knowledge Map ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
Learning Classifier ◽  
Virtual Simulations

Historically, the artificial intelligence (AI) of interactive virtual simulations or games is usually driven by pre-defined static scripts. One of the disadvantages of such scripted opponents is that they can be deciphered and countered by an intelligent user. Thus, the user has the opportunity to find weaknesses and an easy solution against the virtual simulation, which diminishes the efficiency aspect of a training session or entertaining value drastically. While randomization can be used to mask the static behaviour of a scripted AI, it is possible to develop much richer solutions by applying Learning Classifier System (LCS) techniques to create agents with intelligent-like behaviors. Learning Classifier Systems are rule-based machine learning techniques that rely on a Genetic Algorithm to discover a knowledge map used to classify an input space into a set of actions. In this paper, we propose the use of an unsupervised machine learning technique called Accuracy-based Learning Classifier Systems (XCS) for adaptable strategy generation that can be used in virtual simulations or games. XCS use a Genetic Algorithm to evolve a knowledge base in the form of rules. The performance and adaptability of the strategies and tactics developed with the XCS is analyzed by facing these against scripted opponents on a real time strategy game. According to our experiments, the rulesets are able to adapt to a wide array of behaviors from its opponents, as opposed to a linear game script, which is limited in its ability to adapt to its environment.

scholarly journals Real-Valued GCS Classifier System

2007 ◽  
Vol 17 (4) ◽  
pp. 539-547 ◽  
Author(s):  
Łukasz Cielecki ◽  
Olgierd Unold
Keyword(s):  
Evolutionary Computation ◽  
Real World ◽  
Learning Classifier Systems ◽  
Classifier Systems ◽  
Context Free Grammar ◽  
New Model ◽  
Classifier System ◽  
Learning Classifier ◽  
Context Free ◽  
Real World Problems

Real-Valued GCS Classifier SystemLearning Classifier Systems (LCSs) have gained increasing interest in the genetic and evolutionary computation literature. Many real-world problems are not conveniently expressed using the ternary representation typically used by LCSs and for such problems an interval-based representation is preferable. A new model of LCSs is introduced to classify realvalued data. The approach applies the continous-valued context-free grammar-based system GCS. In order to handle data effectively, the terminal rules were replaced by the so-called environment probing rules. The rGCS model was tested on the checkerboard problem.

scholarly journals Self-adaptation of parameters in a learning classifier system ensemble machine

2010 ◽  
Vol 20 (1) ◽  
pp. 157-174 ◽  
Author(s):  
Maciej Troć ◽  
Olgierd Unold
Keyword(s):  
Classifier Systems ◽  
Learning Classifier System ◽  
Classifier System ◽  
System A ◽  
Learning Classifier ◽  
Ensemble Machine Learning ◽  
One Step ◽  
Self Adaptation ◽  
Selection Of ◽  
Self Adaptive

Self-adaptation of parameters in a learning classifier system ensemble machineSelf-adaptation is a key feature of evolutionary algorithms (EAs). Although EAs have been used successfully to solve a wide variety of problems, the performance of this technique depends heavily on the selection of the EA parameters. Moreover, the process of setting such parameters is considered a time-consuming task. Several research works have tried to deal with this problem; however, the construction of algorithms letting the parameters adapt themselves to the problem is a critical and open problem of EAs. This work proposes a novel ensemble machine learning method that is able to learn rules, solve problems in a parallel way and adapt parameters used by its components. A self-adaptive ensemble machine consists of simultaneously working extended classifier systems (XCSs). The proposed ensemble machine may be treated as a meta classifier system. A new self-adaptive XCS-based ensemble machine was compared with two other XCS-based ensembles in relation to one-step binary problems: Multiplexer, One Counts, Hidden Parity, and randomly generated Boolean functions, in a noisy version as well. Results of the experiments have shown the ability of the model to adapt the mutation rate and the tournament size. The results are analyzed in detail.

scholarly journals For Real! XCS with Continuous-Valued Inputs

2003 ◽  
Vol 11 (3) ◽  
pp. 299-336 ◽  
Author(s):  
Christopher Stone ◽  
Larry Bull
Keyword(s):  
Real World ◽  
Sampling Bias ◽  
Learning Classifier Systems ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
System Architectures ◽  
The Real ◽  
Learning Classifier ◽  
Real World Problem ◽  
Associated Operators

Many real-world problems are not conveniently expressed using the ternary representation typically used by Learning Classifier Systems and for such problems an interval-based representation is preferable. We analyse two interval-based representations recently proposed for XCS, together with their associated operators and find evidence of considerable representational and operator bias. We propose a new interval-based representation that is more straightforward than the previous ones and analyse its bias. The representations presented and their analysis are also applicable to other Learning Classifier System architectures. We discuss limitations of the real multiplexer problem, a benchmark problem used for Learning Classifier Systems that have a continuous-valued representation, and propose a new test problem, the checkerboard problem, that matches many classes of real-world problem more closely than the real multiplexer. Representations and operators are compared using both the real multiplexer and checkerboard problems and we find that representational, operator and sampling bias all affect the performance of XCS in continuous-valued environments.

scholarly journals Is Gradient Descent Update Consistent with Accuracy-Based Learning Classifier System?

2009 ◽  
Vol 13 (6) ◽  
pp. 640-648
Author(s):  
Atsushi Wada ◽  
◽  
Keiki Takadama ◽  
◽  
Keyword(s):  
Adaptive Systems ◽  
Gradient Descent ◽  
Benchmark Problems ◽  
Gradient Term ◽  
Rule Discovery ◽  
Discovery Process ◽  
Classifier Systems ◽  
Learning Classifier System ◽  
Classifier System ◽  
Learning Classifier

Learning Classifier Systems (LCSs) are rule-based adaptive systems that have both Reinforcement Learning (RL) and rule-discovery mechanisms for effective and practical online learning. An analysis of the reinforcement process of XCS, one of the currently mainstream LCSs, is performed from the aspect of RL. Upon comparing XCS's update method with gradient-descent-based parameter update in RL, differences are found in the following elements: (1) residual term, (2) gradient term, and (3) payoff definition. All possible combinations of the variants in each element are implemented and tested on multi-step benchmark problems. This revealed that few specific combinations work effectively with XCS's accuracy-based rule-discovery process, while pure gradient-descent-based update showed the worst performance.

Sign in / Sign up

Export Citation Format

Share Document