# Evolution of control with learning classifier systems

@article{Karlsen2018EvolutionOC, title={Evolution of control with learning classifier systems}, author={Matthew R. Karlsen and Sotiris K. Moschoyiannis}, journal={Applied Network Science}, year={2018}, volume={3} }

In this paper we describe the application of a learning classifier system (LCS) variant known as the eXtended classifier system (XCS) to evolve a set of ‘control rules’ for a number of Boolean network instances. We show that (1) it is possible to take the system to an attractor, from any given state, by applying a set of ‘control rules’ consisting of ternary conditions strings (i.e. each condition component in the rule has three possible states; 0, 1 or #) with associated bit-flip actions, and… Expand

#### Figures, Tables, and Topics from this paper

#### 14 Citations

Learning versus optimal intervention in random Boolean networks

- Computer Science
- Appl. Netw. Sci.
- 2019

This paper presents an algorithm for uncovering the optimal set of control rules for controlling random Boolean networks, and assigns relative costs for interventions and ‘natural’ steps, and compares the performance of this optimal rule calculator algorithm (ORC) and the XCS variant of learning classifier systems. Expand

Optimal Control Rules for Random Boolean Networks

- Computer Science
- COMPLEX NETWORKS
- 2018

The performance of an LCS (XCS) on the RBN control problem is assessed in light of the newly uncovered optimal rule set. Expand

Learning to Control Random Boolean Networks: A Deep Reinforcement Learning Approach

- Computer Science
- COMPLEX NETWORKS
- 2019

This paper describes the application of a Deep Reinforcement Learning agent to the problem of control of Gene Regulatory Networks (GRNs) and implements a Deep Q Network with Double Q Learning that is trained by sampling experiences from the environment using Prioritized Experience Replay. Expand

Fine Tuning Run Parameter Values in Rule-Based Machine Learning

- Computer Science
- RuleML+RR
- 2019

This paper investigates the run parameters of Learning Classifier Systems as applied to single-step problems, and studies two LCS variants, XCS for reinforcement learning and UCS for supervised learning, and examines the effect that different parameter values have on enhancing the model prediction, increasing accuracy and reducing the resulting rule set size. Expand

Visualizations for rule-based machine learning

- 2021

Learning Classifier Systems (LCSs) are a group of rule-based evolutionary computation techniques, which have been frequently applied to data mining tasks. The LCSs’ rules are designed to be… Expand

An overview of LCS research from IWLCS 2019 to 2020

- Computer Science
- GECCO Companion
- 2020

An overview of all the LCS-related publications since the last IWLCS, that is, since 13th July 2019 and until 10th March 2020, to provide an easy entry point to the most recent progress and achievements. Expand

Learning Condition-Action Rules for Personalised Journey Recommendations

- Computer Science
- RuleML+RR
- 2018

It is shown that XCSI successfully learns individual passenger preferences and can be used to suggest personalised adjustments to the onward journey in the event of disruption. Expand

Analysis of a network’s asymptotic behavior via its structure involving its strongly connected components

- Computer Science
- Network Science
- 2019

The contributed theorems generalize the existing theoresms on the relation between network structure and asymptotic network behavior addressing only specific cases such as acyclic networks, fully, and strongly connected networks, and theorem addressing only linear functions. Expand

Enabling Serverless Deployment of Large-Scale AI Workloads

- Computer Science
- IEEE Access
- 2020

We propose a set of optimization techniques for transforming a generic AI codebase so that it can be successfully deployed to a restricted serverless environment, without compromising capability or… Expand

Analysis of a Network’s Emerging Behaviour via Its Structure Involving Its Strongly Connected Components

- Computer Science
- Network-Oriented Modeling for Adaptive Networks: Designing Higher-Order Adaptive Biological, Mental and Social Network Models
- 2019

This chapter addressed how network structure can be related to network behaviour and general theorems are presented that relate emerging behaviour of a network to the network’s structure characteristics. Expand

#### References

SHOWING 1-10 OF 82 REFERENCES

Classifier Fitness Based on Accuracy

- Mathematics, Computer Science
- Evolutionary Computation
- 1995

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. Expand

Discrete dynamical genetic programming in XCS

- Computer Science
- GECCO
- 2009

Results are presented from an investigation into using a discrete dynamical system representation within the XCS Learning Classifier System, where asynchronous random Boolean networks are used to represent the traditional condition-action production system rules. Expand

The stability of long action chains in XCS

- Mathematics, Computer Science
- Soft Comput.
- 2002

It is shown that the introduction of generalisation introduces low bounds on the length of action-chains that can be identified and chosen between to find the optimal pathway, and the Domination Hypothesis is introduced to explain its formation and preservation. Expand

On dynamical genetic programming: simple Boolean networks in learning classifier systems

- Computer Science
- Int. J. Parallel Emergent Distributed Syst.
- 2009

It is shown possible to evolve ensembles of dynamical Boolean function networks to solve versions of the well-known multiplexer problem. Expand

An Analysis of Generalization in the XCS Classifier System

- Mathematics, Computer Science
- Evolutionary Computation
- 1999

It is shown that XCS's generalization mechanism is effective, but that the conditions under which it works must be clearly understood, and the compactness of the representation evolved by XCS is limited by the number of instances of each generalization actually present in the environment. Expand

Gradient descent methods in learning classifier systems: improving XCS performance in multistep problems

- Mathematics, Computer Science
- IEEE Transactions on Evolutionary Computation
- 2005

The extension of XCS to gradient-based update methods results in a classifier system that is more robust and more parameter independent, solving large and difficult maze problems reliably. Expand

Learning classifier systems: a complete introduction, review, and roadmap

- Computer Science
- 2009

This paper aims to provide an accessible foundation for researchers of different backgrounds interested in selecting or developing their own LCS, including a simple yet thorough introduction, a historical review, and a roadmap of algorithmic components, emphasizing differences in alternative LCS implementations. Expand

External Control in Markovian Genetic Regulatory Networks

- Mathematics, Computer Science
- Machine Learning
- 2004

This paper develops a procedure by which one can choose the sequence of control actions that minimize a given performance index over a finite number of steps, based on the theory of controlled Markov chains and makes use of the classical technique of Dynamic Programming. Expand

Emergent behavior in classifier systems

- Computer Science, Mathematics
- 1990

The mapping provides a way to understand and predict emergent classifier system behaviors by observing the dynamical behavior of the Boolean networks. Expand

An introduction to genetic algorithms

- Computer Science
- 1996

An Introduction to Genetic Algorithms focuses in depth on a small set of important and interesting topics -- particularly in machine learning, scientific modeling, and artificial life -- and reviews a broad span of research, including the work of Mitchell and her colleagues. Expand