Filters








2,769 Hits in 3.5 sec

Revisit of Rule-Deletion Strategy for XCSAM Classifier System on Classification

Masaya Nakata, Tomoki Hamagami
2017 Transactions of the Institute of Systems Control and Information Engineers  
Different from the existing deletion strategy which deletes two rules for each rule-evolution, our deletion strategy is modified to delete more than two rules as necessary.  ...  This paper starts with claiming a limitation that XCSAM still fails to evolutionary generate adequate rules advocating the highest-return actions.  ...  Then, rather than building the pure best action map, a new concept of learning strategy which combines best action map with complete action map, may be a good solution to overcome the limitations of both  ... 
doi:10.5687/iscie.30.273 fatcat:eba2jzyvszdabfeqwyllsflp44

Active Reinforcement Learning – A Roadmap Towards Curious Classifier Systems for Self-Adaptation [article]

Simon Reichhuber, Sven Tomforde
2022 arXiv   pre-print
In this context, the fundamental reinforcement learning approaches come with several drawbacks that hinder their application to real-world systems: trial-and-error, purely reactive behaviour or isolated  ...  The idea of this article is to present a concept for alleviating these drawbacks by setting up a research agenda towards what we call "active reinforcement learning" in intelligent systems.  ...  Figure 6 : 6 Figure 6: Self-adaptive XCS with observer/controller tasks.  ... 
arXiv:2201.03947v1 fatcat:6dgept56bfg2bdiqvvlrlm6ha4

Evolutionary Online Data Mining: An Investigation in a Dynamic Environment [chapter]

Hai H. Dam, Chris Lokan, Hussein A. Abbass
2007 Studies in Computational Intelligence  
Online or incremental learning becomes more important than ever for dealing with these problems.  ...  We propose several strategies to force the system to learn quickly after severe changes. There are adaptive learning; re-initializing the parameters; and re-initializing the population.  ...  During the selection process, two parents from the action set [A] are selected with a probability proportional to their fitness.  ... 
doi:10.1007/978-3-540-49774-5_7 fatcat:ranrejqwtndn3ou7ubdyti36oq

Special issue on the 20th anniversary of XCS

Tim Kovacs, Muhammad Iqbal, Kamran Shafi, Ryan Urbanowicz
2015 Evolutionary Intelligence  
[9] introduced XCSAM (XCS with Adaptive action Mapping), which estimates how often each rule takes an optimal action and incorporates that estimate into the rule's fitness.  ...  They then introduce a new variant of XCSAM with a new selection strategy which results in smaller population sizes.  ... 
doi:10.1007/s12065-015-0131-0 fatcat:2f52bxzpzfd77hz3je26azrtei

Learning from Data using XCS

Elias B. Ayele, Abdollah Homaifar, Albert Esterline, Robert Dean, Dan Rodgers
2008 IFAC Proceedings Volumes  
We detail the TRWA and the significant improvements made to the usual XCS strategies in order to achieve our goal of using a supervised learning technique for the TRWA.  ...  The use of tournament selection instead of roulette wheel selection for selecting two parents in the GA is also analyzed.  ...  Various input-output mappings are provided as part of the training examples, and the learner is given some sort of reinforcement for selecting the best output for a given input.  ... 
doi:10.3182/20080706-5-kr-1001.02613 fatcat:2d7qotypazcjzbqss2wl6sujqi

Evaluating the XCS learning classifier system in competitive simultaneous learning environments

Neera P. Sood, Ashley G. Williams, Kenneth A. De Jong
2005 Proceedings of the 2005 workshops on Genetic and evolutionary computation - GECCO '05  
We are interested in seeing whether it can offer an accessible representation model and evolve feasible strategies to predict future demand patterns endogenously, and in parallel with the supply side simulation  ...  We would like to evaluate the XCS [1] Learning Classifier System (LCS [2]) to see if it can be applied to a specific aviation industry problem.  ...  ACKNOWLEDGMENTS We gratefully acknowledge the assistance and support provided by The MITRE Corporation for this research. We are especially grateful to Ashley G.  ... 
doi:10.1145/1102256.1102282 dblp:conf/gecco/SoodWJ05 fatcat:t3wt36to3netdcgyzdgojlrd2y

An Empirical Analysis of Action Map in Learning Classifier Systems

Masaya NAKATA, Keiki TAKADAMA
2018 SICE Journal of Control Measurement and System Integration  
This paper attempts to empirically conclude this issue with an intensive analysis comparing different action maps on LCSs.  ...  Those results claim that the action map should be selected adequately in designing LCSs in order to improve their potential performance.  ...  Then, the best action map potentially can be a better strategy than the complete action map.  ... 
doi:10.9746/jcmsi.11.239 fatcat:y4hnbkw34jaazlowfa22oltjta

Self-learning Smart Cameras - Harnessing the Generalization Capability of XCS

Anthony Stein, Stefan Rudolph, Sven Tomforde, Jörg Hähner
2017 Proceedings of the 9th International Joint Conference on Computational Intelligence  
More precisely, the Extended Classifier System (XCS) is utilized to learn a configuration strategy for the pan, tilt, and zoom of smart cameras.  ...  which is based on Q-Learning, by harnessing the generalization capability of Learning Classifier Systems (LCS), i.e. avoiding to separately approximate the quality of each possible (re-)configuration (action  ...  Regarding the exploration/exploitation trade-off, we utilized the ε-greedy action-selection strategy, where ε determines the probability of choosing a random action instead of the action with the maximum  ... 
doi:10.5220/0006512101290140 dblp:conf/ijcci/SteinRTH17 fatcat:fixmlcuvxvhfnk57aso2hmhgfm

A novel adaptive pigeon-inspired optimization algorithm based on evolutionary game theory

Xingshuo Hai, Zili Wang, Qiang Feng, Yi Ren, Bo Sun, Dezhen Yang
2020 Science China Information Sciences  
A novel adaptive pigeon-inspired optimization algorithm based on evolutionary game theory. Sci China Inf Sci, 2021, 64(3): 139203, https://doi.  ...  For a player, the payoff consists of the cost and the ratio of its strategy.  ...  Therefore, it is possible to automatically select the weighting coefficients of PIO in accordance with the specific problem instead of choosing them within the experiential interval.  ... 
doi:10.1007/s11432-018-9923-6 fatcat:ctdwpbezavgx7c72ycx6lpsgxa

A multiple population XCS: Evolving condition-action rules based on feature space partitions

Mani Abedini, Michael Kirley
2010 IEEE Congress on Evolutionary Computation  
The behavior of the multiple population model is carefully analyzed and compared with the original XCS using the Boolean logic multiplexer problem as a test case.  ...  XCS is an accuracy-based machine learning technique, which combines reinforcement learning and evolutionary algorithms to evolve a set of classifiers (or rules) for pattern classification tasks.  ...  strategy with migration -in CoXCS.  ... 
doi:10.1109/cec.2010.5586521 dblp:conf/cec/AbediniK10 fatcat:nkwqm5lfozch5osq4yzkkjxidy

Rule Fitness and Pathology in Learning Classifier Systems

Tim Kovacs
2004 Evolutionary Computation  
We begin with the basics, considering definitions for correct and incorrect actions, and then correct, incorrect, and overgeneral rules for both strength and accuracy-based fitness.  ...  Next we show how to design fit overgeneral rules for XCS (but not SB-XCS), by introducing biases in the variance of the reward function, and thus that each system has its own weakness.  ...  for action selection.  ... 
doi:10.1162/106365604773644341 pmid:15096307 fatcat:amt7ba5io5dp3nx6psfiokf24y

Rule Fitness and Pathology in Learning Classifier Systems

Tim Kovacs
2004 Evolutionary Computation  
We begin with the basics, considering definitions for correct and incorrect actions, and then correct, incorrect, and overgeneral rules for both strength and accuracy-based fitness.  ...  Next we show how to design fit overgeneral rules for XCS (but not SB-XCS), by introducing biases in the variance of the reward function, and thus that each system has its own weakness.  ...  for action selection.  ... 
doi:10.1162/evco.2004.12.1.99 pmid:15096307 fatcat:4th7w4gggffevkv73gaj2dvw2u

Using the XCS Classifier System for Multi-objective Reinforcement Learning Problems

Matthew Studley, Larry Bull
2007 Artificial Life  
Investigations of how XCS performs in other types of multi-objective learning tasks are also being undertaken, along with comparisons with other approaches [1] .  ...  We are now exploiting these findings to use XCS to control a real mobile robot that must solve such multi-objective problems.  ...  In Figure 19 we see the effect of N on the performance of XCS in the Woods1ef problem, using the two exploration action-selection strategies.  ... 
doi:10.1162/artl.2007.13.1.69 pmid:17204013 fatcat:zhcq7z3csvbypnaksjb2e7shda

Tournament Selection: Stable Fitness Pressure in XCS [chapter]

Martin V. Butz, Kumara Sastry, David E. Goldberg
2003 Lecture Notes in Computer Science  
Also in the accuracy-based learning classifier system XCS, introduced by Wilson in 1995, proportionate selection is used.  ...  Consequently, tournament selection is introduced which makes XCS more parameter independent, noise independent, and more efficient in exploiting fitness guidance.  ...  Additional support from the Automated Learning Group (ALG) at the National Center for Supercomputing Applications (NCSA) is acknowledged.  ... 
doi:10.1007/3-540-45110-2_83 fatcat:nh57qktl7ng6zblymqkoxr5oey

Learning Classifier Systems: A Complete Introduction, Review, and Roadmap

Ryan J. Urbanowicz, Jason H. Moore
2009 Journal of Artificial Evolution and Applications  
This paper aims to provide an accessible foundation for researchers of different backgrounds interested in selecting or developing their own LCS.  ...  The LCS concept has inspired a multitude of implementations adapted to manage the different problem domains to which it has been applied (e.g., autonomous robotics, classification, knowledge discovery,  ...  The complete action mapping of XCS made it possible to address the problem of function approximation.  ... 
doi:10.1155/2009/736398 fatcat:dvubeerbl5ez3ed553sn2snofy
« Previous Showing results 1 — 15 out of 2,769 results