Category: Uncategorized

Python Code for AF_UCS_2.0 with Multicore Parallelization Now Available

AF_UCS_2.0_Multicore_Parallelization

We have organized, annotated, and cleaned up the code for our published Michigan-Style Learning Classifier System implementations.  AF_UCS stands for Attribute Feedback Supervised Classifier System.  The above code was utilized in the following publications:

  1. Rudd, J., Moore, JH., Urbanowicz, R. A simple multi-core parallelization strategy for learning classifier system evaluation. Proceeding of the fifteenth annual conference companion on Genetic and evolutionary computation conference companion. ACM. 1259-1266, 2013
  2. Rudd, J., Moore, JH., Urbanowicz, R. A multi-core parallelization strategy for statistical significance testing in learning classifier systems. Evolutionary Intelligence. {in press} 2013

Python Code for EK_AF_UCS_1.0 Now Available

EK_AF_UCS_1.0

We have organized, annotated, and cleaned up the code for our published Michigan-Style Learning Classifier System implementations.  EK_AF_UCS stands for Expert Knowledge and Attribute Feedback Supervised Classifier System.  The above code was utilized in the following publications:

  1. Urbanowicz, R., Andrew, A., Karagas, M., Moore, J. Role of genetic heterogeneity and epistasis in bladder cancer susceptibility and outcome: a learning classifier system approach.  Journal of the American Medical Informatics Association. 20:4, 603-612, 2013
  2. Urbanowicz, R., Granizo-Mackenzie, D., Moore, J. Using expert knowledge to guide covering and mutation in a michigan style learning classifier system to detect epistasis and heterogeneity.  Parallel Problem Solving from Nature-PPSN XII. Springer. 266-275, 2012

Python Code for AF_UCS_1.0 Now Available

AF_UCS_1.0

We have organized, annotated, and cleaned up the code for our published Michigan-Style Learning Classifier System implementations.  AF_UCS stands for Attribute Feedback Supervised Classifier System.  The above code was utilized in the following publications:

  1. Urbanowicz, R., Granizo-Mackenzie, A., Moore, J. An Analysis Pipeline with Statistical and Visualization-Guided Knowledge Discovery for Michigan-Style Learning Classifier Systems. Computational Intelligence Magazine, IEEE. 35-45, 2012
  2. Urbanowicz, R., Granizo Mackenzie, A., Moore, J. Instance-linked attribute tracking and feedback for michigan-style supervised learning classifier systems. Proceedings of the fourteenth international conference on Genetic and evolutionary computation conference. ACM. 927-934, 2012

Classification-based self-adaptive differential evolution with fast and reliable convergence performance

Abstract  To avoid the problems of slow and premature convergence of the differential evolution (DE) algorithm, this paper presents
a new DE variant named p-ADE. It improves the convergence performance by implementing a new mutation strategy…

Abstract  

To avoid the problems of slow and premature convergence of the differential evolution (DE) algorithm, this paper presents
a new DE variant named p-ADE. It improves the convergence performance by implementing a new mutation strategy “DE/rand-to-best/pbest”,
together with a classification mechanism, and controlling the parameters in a dynamic adaptive manner, where the “DE/rand-to-best/pbest”
utilizes the current best solution together with the best previous solution of each individual to guide the search direction.
The classification mechanism helps to balance the exploration and exploitation of individuals with different fitness characteristics,
thus improving the convergence rate. Dynamic self-adaptation is beneficial for controlling the extent of variation for each
individual. Also, it avoids the requirement for prior knowledge about parameter settings. Experimental results confirm the
superiority of p-ADE over several existing DE variants as well as other significant evolutionary optimizers.

  • Content Type Journal Article
  • Pages 1-19
  • DOI 10.1007/s00500-010-0689-5
  • Authors
    • Xiao-Jun Bi, Department of Information and Communication Engineering, Harbin Engineering University, Harbin, 150001 China
    • Jing Xiao, Department of Information and Communication Engineering, Harbin Engineering University, Harbin, 150001 China

A fuzzy-rule-based driving architecture for non-player characters in a car racing game

Abstract  Videogame-based competitions have been the target of considerable interest among researchers over the past few years since
they provide an ideal framework in which to apply soft computing techniques. One of the most popular competi…

Abstract  

Videogame-based competitions have been the target of considerable interest among researchers over the past few years since
they provide an ideal framework in which to apply soft computing techniques. One of the most popular competitions is the Simulated Car Racing Competition which, thanks to the realism implemented by recent car simulators, provides an excellent test bed for the application of
autonomous driving techniques. The present work describes the design and implementation of a car controller able to deal with
competitive racing situations. The complete driving architecture consists of six simple modules, each one responsible for
a basic aspect of car driving. Three modules use simple functions to control gear shifting, steering movements, and pedal
positions. A fourth manages speed control by means of a simple fuzzy system. The other two modules are in charge of (i) adapting
the driving behaviour to the presence of other cars, and (ii) implementing a basic ‘inter-lap’ learning mechanism in order
to remember key track segments and adapt the speed accordingly in future laps. The controller was evaluated in two ways. First,
in runs without adversaries over several track designs, our controller allowed some of the longest distances to be covered
in a set time in comparison with data from other previous controllers, and second, as a participant in the 2009 Simulated Car Racing Competition which it ended up winning.

  • Content Type Journal Article
  • Pages 1-13
  • DOI 10.1007/s00500-011-0691-6
  • Authors
    • Enrique Onieva, AUTOPIA program of the Center for Automation and Robotics, Universidad Politécnica de Madrid–Consejo Superior de Investigaciones Científicas, La Poveda-Arganda del Rey, 28500 Madrid, Spain
    • David A. Pelta, Department of Computer Science and Artificial Intelligence, University of Granada, 18071 Granada, Spain
    • Vicente Milanés, AUTOPIA program of the Center for Automation and Robotics, Universidad Politécnica de Madrid–Consejo Superior de Investigaciones Científicas, La Poveda-Arganda del Rey, 28500 Madrid, Spain
    • Joshue Pérez, AUTOPIA program of the Center for Automation and Robotics, Universidad Politécnica de Madrid–Consejo Superior de Investigaciones Científicas, La Poveda-Arganda del Rey, 28500 Madrid, Spain

Introduction

Introduction
Content Type Journal ArticlePages 1-3DOI 10.1007/s11047-011-9247-zAuthors
José Félix Costa, Department of Mathematics, Instituto Superior Técnico Universidade Técnica de Lisboa, Lisboa, PortugalNachum Dershowitz, School of Computer …

Introduction

  • Content Type Journal Article
  • Pages 1-3
  • DOI 10.1007/s11047-011-9247-z
  • Authors
    • José Félix Costa, Department of Mathematics, Instituto Superior Técnico Universidade Técnica de Lisboa, Lisboa, Portugal
    • Nachum Dershowitz, School of Computer Science, Tel Aviv University, Ramat Aviv, Israel

Genetic programming with one-point crossover and subtree mutation for effective problem solving and bloat control

Abstract  Genetic programming (GP) is one of the most widely used paradigms of evolutionary computation due to its ability to automatically
synthesize computer programs and mathematical expressions. However, because GP uses a variable length…

Abstract  

Genetic programming (GP) is one of the most widely used paradigms of evolutionary computation due to its ability to automatically
synthesize computer programs and mathematical expressions. However, because GP uses a variable length representation, the
individuals within the evolving population tend to grow rapidly without a corresponding return in fitness improvement, a phenomenon
known as bloat. In this paper, we present a simple bloat control strategy for standard tree-based GP that achieves a one order
of magnitude reduction in bloat when compared with standard GP on benchmark tests, and practically eliminates bloat on two
real-world problems. Our proposal is to substitute standard subtree crossover with the one-point crossover (OPX) developed
by Poli and Langdon (Second online world conference on soft computing in engineering design and manufacturing, Springer, Berlin
(1997)), while maintaining all other GP aspects standard, particularly subtree mutation. OPX was proposed for theoretical purposes
related to GP schema theorems, however since it curtails exploration during the search it has never achieved widespread use.
In our results, on the other hand, we are able to show that OPX can indeed perform an effective search if it is coupled with
subtree mutation, thus combining the bloat control capabilities of OPX with the exploration provided by standard mutation.

  • Content Type Journal Article
  • Pages 1-17
  • DOI 10.1007/s00500-010-0687-7
  • Authors
    • Leonardo Trujillo, Instituto Tecnológico de Tijuana, Av. Tecnológico S/N, Fracc. Tomás Aquino, Tijuana, BC, Mexico

Use of the q-Gaussian mutation in evolutionary algorithms

Abstract  This paper proposes the use of the q-Gaussian mutation with self-adaptation of the shape of the mutation distribution in evolutionary algorithms. The shape of
the q-Gaussian mutation distribution is controlled by a real parameter q…

Abstract  

This paper proposes the use of the q-Gaussian mutation with self-adaptation of the shape of the mutation distribution in evolutionary algorithms. The shape of
the q-Gaussian mutation distribution is controlled by a real parameter q. In the proposed method, the real parameter q of the q-Gaussian mutation is encoded in the chromosome of individuals and hence is allowed to evolve during the evolutionary process.
In order to test the new mutation operator, evolution strategy and evolutionary programming algorithms with self-adapted q-Gaussian mutation generated from anisotropic and isotropic distributions are presented. The theoretical analysis of the q-Gaussian mutation is also provided. In the experimental study, the q-Gaussian mutation is compared to Gaussian and Cauchy mutations in the optimization of a set of test functions. Experimental
results show the efficiency of the proposed method of self-adapting the mutation distribution in evolutionary algorithms.

  • Content Type Journal Article
  • Pages 1-27
  • DOI 10.1007/s00500-010-0686-8
  • Authors
    • Renato Tinós, Department of Physics and Mathematics, FFCLRP, University of São Paulo (USP), Ribeirão Preto, SP 14040-901, Brazil
    • Shengxiang Yang, Department of Information Systems and Computing, Brunel University, Uxbridge, Middlesex UB8 3PH, UK