Author: Community

Evolutionary parallel and gradually distributed lateral tuning of fuzzy rule-based systems

Abstract  The tuning of Fuzzy Rule-Based Systems is often applied to improve their performance as a post-processing stage once an initial
set of fuzzy rules has been extracted. This optimization problem can become a hard one when the size of…

Abstract  The tuning of Fuzzy Rule-Based Systems is often applied to improve their performance as a post-processing stage once an initial
set of fuzzy rules has been extracted. This optimization problem can become a hard one when the size of the considered system
in terms of the number of variables, rules and, particularly, data samples is big. Distributed Genetic Algorithms are excellent
optimization algorithms which exploit the nowadays available parallel hardware (multicore microprocessors and clusters) and
could help to alleviate this growth in complexity. In this work, we present a study on the use of the Distributed Genetic
Algorithms for the tuning of Fuzzy Rule-Based Systems. To this end, we analyze the application of a specific Gradual Distributed
Real-Coded Genetic Algorithm which employs eight subpopulations in a hypercube topology and local parallelization at each
subpopulation. We tested our approach on nine real-world datasets of different sizes and with different numbers of variables.
The empirical performance in solution quality and computing time is assessed by comparing its results with those from a highly
effective sequential tuning algorithm. The results show that the distributed approach achieves better results in terms of
quality and execution time as the complexity of the problem grows.

  • Content Type Journal Article
  • DOI 10.1007/s12065-009-0025-0
  • Authors
    • I. Robles, University of Granada Dept. of Computer Sciences and Artificial Intelligence 18071 Granada Spain
    • R. Alcalá, University of Granada Dept. of Computer Sciences and Artificial Intelligence 18071 Granada Spain
    • J. M. Benítez, University of Granada Dept. of Computer Sciences and Artificial Intelligence 18071 Granada Spain
    • F. Herrera, University of Granada Dept. of Computer Sciences and Artificial Intelligence 18071 Granada Spain

Multi-objective evolutionary learning of granularity, membership function parameters and rules of Mamdani fuzzy systems

Abstract  In this paper, we propose a multi-objective evolutionary algorithm (MOEA) to generate Mamdani fuzzy rule-based systems with
different trade-offs between accuracy and complexity by learning concurrently granularities of the input an…

Abstract  In this paper, we propose a multi-objective evolutionary algorithm (MOEA) to generate Mamdani fuzzy rule-based systems with
different trade-offs between accuracy and complexity by learning concurrently granularities of the input and output partitions,
membership function (MF) parameters and rules. To this aim, we introduce the concept of virtual and concrete partitions: the
former is defined by uniformly partitioning each linguistic variable with a fixed maximum number of fuzzy sets; the latter
takes into account, for each variable, the number of fuzzy sets determined by the evolutionary process. Rule bases and MF
parameters are defined on the virtual partitions and, whenever a fitness evaluation is required, mapped to the concrete partitions
by employing appropriate mapping strategies. The implementation of the MOEA relies on a chromosome composed of three parts,
which codify the partition granularities, the virtual rule base and the membership function parameters, respectively, and
on purposely-defined genetic operators. The MOEA has been tested on three real-world regression problems achieving very promising
results. In particular, we highlight how starting from randomly generated solutions, the MOEA is able to determine different
granularities for different variables achieving good trade-offs between complexity and accuracy.

  • Content Type Journal Article
  • DOI 10.1007/s12065-009-0022-3
  • Authors
    • Michela Antonelli, University of Pisa Dipartimento di Ingegneria dell’Informazione: Elettronica, Informatica, Telecomunicazioni Via Diotisalvi 2 56122 Pisa Italy
    • Pietro Ducange, University of Pisa Dipartimento di Ingegneria dell’Informazione: Elettronica, Informatica, Telecomunicazioni Via Diotisalvi 2 56122 Pisa Italy
    • Beatrice Lazzerini, University of Pisa Dipartimento di Ingegneria dell’Informazione: Elettronica, Informatica, Telecomunicazioni Via Diotisalvi 2 56122 Pisa Italy
    • Francesco Marcelloni, University of Pisa Dipartimento di Ingegneria dell’Informazione: Elettronica, Informatica, Telecomunicazioni Via Diotisalvi 2 56122 Pisa Italy

Hybrid of genetic algorithm and local search to solve MAX-SAT problem using nVidia CUDA framework

Abstract  General Purpose computing over Graphical Processing Units (GPGPUs) is a huge shift of paradigm in parallel computing that
promises a dramatic increase in performance. But GPGPUs also bring an unprecedented level of complexity in al…

Abstract  General Purpose computing over Graphical Processing Units (GPGPUs) is a huge shift of paradigm in parallel computing that
promises a dramatic increase in performance. But GPGPUs also bring an unprecedented level of complexity in algorithmic design
and software development. In this paper we describe the challenges and design choices involved in parallelizing a hybrid of
Genetic Algorithm (GA) and Local Search (LS) to solve MAXimum SATisfiability (MAX-SAT) problem on a state-of-the-art nVidia
Tesla GPU using nVidia Compute Unified Device Architecture (CUDA). MAX-SAT is a problem of practical importance and is often
solved by employing metaheuristics based search methods like GAs and hybrid of GA with LS. Almost all the parallel GAs (pGAs)
designed in the last two decades were designed for either clusters or MPPs. Unfortunately, very little research is done on
the implementation of such algorithms over commodity graphics hardware. GAs in their simple form are not suitable for implementation
over the Single Instruction Multiple Thread (SIMT) architecture of a GPU, and the same is the case with conventional LS algorithms.
In this paper we explore different genetic operators that can be used for an efficient implementation of GAs over nVidia GPUs.
We also design and introduce new techniques/operators for an efficient implementation of GAs and LS over such architectures.
We use nVidia Tesla C1060 to perform several numerical tests and performance measurements and show that in the best case we
obtain a speedup of 25×. We also discuss the effects of different optimization techniques on the overall execution time.

  • Content Type Journal Article
  • Pages 391-415
  • DOI 10.1007/s10710-009-9091-4
  • Authors
    • Asim Munawar, Hokkaido University Graduate School of Information Science & Technology Sapporo Japan
    • Mohamed Wahib, Hokkaido University Graduate School of Information Science & Technology Sapporo Japan
    • Masaharu Munetomo, Hokkaido University Information Initiative Center Sapporo Japan
    • Kiyoshi Akama, Hokkaido University Information Initiative Center Sapporo Japan

Parallel evolution using multi-chromosome cartesian genetic programming

Abstract  Parallel and distributed methods for evolutionary algorithms have concentrated on maintaining multiple populations of genotypes,
where each genotype in a population encodes a potential solution to the problem. In this paper, we inv…

Abstract  Parallel and distributed methods for evolutionary algorithms have concentrated on maintaining multiple populations of genotypes,
where each genotype in a population encodes a potential solution to the problem. In this paper, we investigate the parallelisation
of the genotype itself into a collection of independent chromosomes which can be evaluated in parallel. We call this multi-chromosomal evolution
(MCE). We test this approach using Cartesian Genetic Programming and apply MCE to a series of digital circuit design problems
to compare the efficacy of MCE with a conventional single chromosome approach (SCE). MCE can be readily used for many digital
circuits because they have multiple outputs. In MCE, an independent chromosome is assigned to each output. When we compare
MCE with SCE we find that MCE allows us to evolve solutions much faster. In addition, in some cases we were able to evolve
solutions with MCE that we unable to with SCE. In a case-study, we investigate how MCE can be applied to to a single objective
problem in the domain of image classification, namely, the classification of breast X-rays for cancer. To apply MCE to this
problem, we identify regions of interest (RoI) from the mammograms, divide the RoI into a collection of sub-images and use
a chromosome to classify each sub-image. This problem allows us to evaluate various evolutionary mutation operators which
can pairwise swap chromosomes either randomly or topographically or reuse chromosomes in place of other chromosomes.

  • Content Type Journal Article
  • Pages 417-445
  • DOI 10.1007/s10710-009-9093-2
  • Authors
    • James Alfred Walker, University of York Intelligent Systems Group, Department of Electronics Heslington York YO10 5DD UK
    • Katharina Völk, University of York Intelligent Systems Group, Department of Electronics Heslington York YO10 5DD UK
    • Stephen L. Smith, University of York Intelligent Systems Group, Department of Electronics Heslington York YO10 5DD UK
    • Julian Francis Miller, University of York Intelligent Systems Group, Department of Electronics Heslington York YO10 5DD UK

On the regularity of circular splicing languages: a survey and new developments

Abstract  
Circular splicing has been introduced to model a specific recombinant behaviour of circular DNA, continuing the investigation initiated with
linear splicing. In this paper we focus on the relationship between regular circular lan…

Abstract  

Circular splicing has been introduced to model a specific recombinant behaviour of circular DNA, continuing the investigation initiated with
linear splicing. In this paper we focus on the relationship between regular circular languages and languages generated by finite circular splicing systems. We survey the known results towards a characterization of the intersection between these two classes and provide new contributions
on the open problem of finding this characterization. First, we exhibit a non-regular circular language generated by a circular simple system thus disproving a known result in this area. Then we give new results related to a restrictive class of circular splicing
systems, the marked systems. Precisely, we review in a graph theoretical setting the recently obtained characterization of marked systems generating
regular circular languages. In particular, we define a slight variant of marked systems, that is the g-marked systems, and we introduce the graph associated with a g-marked system. We show that a g-marked system generates a regular circular
language if and only if its associated graph is a cograph. Furthermore, we prove that the class of g-marked systems generating regular circular languages is closed under a complement
operation applied to systems. We also prove that marked systems with self-splicing generate only regular circular languages.

  • Content Type Journal Article
  • DOI 10.1007/s11047-009-9155-7
  • Authors
    • Paola Bonizzoni, Università degli Studi di Milano-Bicocca Dipartimento di Informatica Sistemistica e Comunicazione Viale Sarca 336 20126 Milano Italy
    • Clelia De Felice, Università di Salerno Dipartimento di Informatica ed Applicazioni 84084 Fisciano SA Italy
    • Gabriele Fici, Università di Salerno Dipartimento di Informatica ed Applicazioni 84084 Fisciano SA Italy
    • Rosalba Zizza, Università di Salerno Dipartimento di Informatica ed Applicazioni 84084 Fisciano SA Italy

Tweet for engineering (& computer science) education

The Alliance for Promoting Innovation in Engineering Education (aPIE2) is launching a new initiative,  starting Monday, 19 October 2009 at the 2009 IEEE Frontiers in Education: Twitter for Engineering Education Transformation and Innovation (TwEETI).
More information on the TwEETI initiative is availabe at the TwEETI tab of the www.apie2.org website, but the idea is to encourage […]

The Alliance for Promoting Innovation in Engineering Education (aPIE2) is launching a new initiative,  starting Monday, 19 October 2009 at the 2009 IEEE Frontiers in Education: Twitter for Engineering Education Transformation and Innovation (TwEETI).

More information on the TwEETI initiative is availabe at the TwEETI tab of the www.apie2.org website, but the idea is to encourage more engineering educators and friends of engineering education to start using twitter to share and vet interesting innovations and deveopments in engineering education.

Existing twitter accounts can join by simply following @apie2.  aPIE2 will itself follow accounts with significant engineering education content or information of interest to educators and friends of engineering education.  Those not on twitter who would like to participate can sign up by following the instructions here.

For more information contact Dave Goldberg (deg@illinois.edu or (@deg511).  @aPIE2, @deg511, and @iFoundry will twitter IEEE FIE live and others attending the conference are invited to join in.  The hash tag for FIE will be #fie09.

Related Posts

    Transducer generated arrays of robotic nano-arms

    Abstract  We consider sets of two-dimensional arrays, called here transducer generated languages, obtained by iterative applications
    of transducers (finite state automata with output). Each transducer generates a set of blocks of symbols suc…

    Abstract  

    We consider sets of two-dimensional arrays, called here transducer generated languages, obtained by iterative applications
    of transducers (finite state automata with output). Each transducer generates a set of blocks of symbols such that the bottom
    row of a block is an input string accepted by the transducer and, by iterative application of the transducer, each row of
    the block is an output of the transducer on the preceding row. We show how these arrays can be implemented through molecular
    assembly of triple crossover DNA molecules. Such assembly could serve as a scaffold for arranging molecular robotic arms capable
    of simultaneous movements. We observe that transducer generated languages define a class of languages which is a proper subclass
    of recognizable picture languages, but it contains the class of all factorial local two-dimensional languages. By taking the
    average growth rate of the number of blocks in the language as a measure of its complexity, we further observe that arrays
    with high complexity patterns can be generated in this way.

    • Content Type Journal Article
    • DOI 10.1007/s11047-009-9157-5
    • Authors
      • Egor Dolzhenko, University of South Florida Department of Mathematics and Statistics Tampa FL 33620 USA
      • Nataša Jonoska, University of South Florida Department of Mathematics and Statistics Tampa FL 33620 USA
      • Nadrian C. Seeman, New York University Chemistry Department New York NY 10003 USA

    Genetic programming on graphics processing units

    Abstract  The availability of low cost powerful parallel graphics cards has stimulated the port of Genetic Programming (GP) on Graphics
    Processing Units (GPUs). Our work focuses on the possibilities offered by Nvidia G80 GPUs when programmed…

    Abstract  The availability of low cost powerful parallel graphics cards has stimulated the port of Genetic Programming (GP) on Graphics
    Processing Units (GPUs). Our work focuses on the possibilities offered by Nvidia G80 GPUs when programmed in the CUDA language.
    In a first work we have showed that this setup allows to develop fine grain parallelization schemes to evaluate several GP
    programs in parallel, while obtaining speedups for usual training sets and program sizes. Here we present another parallelization
    scheme and optimizations about program representation and use of GPU fast memory. This increases the computation speed about
    three times faster, up to 4 billion GP operations per second. The code has been developed within the well known ECJ library
    and is open source.

    • Content Type Journal Article
    • Pages 447-471
    • DOI 10.1007/s10710-009-9092-3
    • Authors
      • Denis Robilliard, Univ. Lille Nord de France, ULCO, LIL Maison de la Recherche Blaise Pascal, 50 rue Ferdinand Buisson BP 719 62228 CALAIS Cedex France
      • Virginie Marion-Poty, Univ. Lille Nord de France, ULCO, LIL Maison de la Recherche Blaise Pascal, 50 rue Ferdinand Buisson BP 719 62228 CALAIS Cedex France
      • Cyril Fonlupt, Univ. Lille Nord de France, ULCO, LIL Maison de la Recherche Blaise Pascal, 50 rue Ferdinand Buisson BP 719 62228 CALAIS Cedex France

    Laura Landweber to present “The evolution of cellular computing: A modern take on Lamarckian inheritance”

    Monday, October 12  7:00 pm
    1320 Digital Computer Lab
    1304 W. Springfield
    2009 marks not only the 200th anniversary of Darwin’s birth but also publication of the first evolutionary theory, Lamarck’s Philosophie Zoologique.  “Lamarckian evolution” is usually distinguished from “Darwinian evolution” by its reliance on the inheritance of acquired characteristics, and as such, is usually dismissed as lacking […]

    Monday, October 12  7:00 pm
    1320 Digital Computer Lab
    1304 W. Springfield

    2009 marks not only the 200th anniversary of Darwin’s birth but also publication of the first evolutionary theory, Lamarck’s Philosophie Zoologique.  “Lamarckian evolution” is usually distinguished from “Darwinian evolution” by its reliance on the inheritance of acquired characteristics, and as such, is usually dismissed as lacking a sound biological basis.  However, there is increasingly compelling evidence for a variety of molecular mechanisms that can support Lamarckian modes of evolution.  In this talk, I will discuss our recent work with pond organisms, known as ciliates, demonstrating an extraordinary new role for the molecule RNA.  Normally thought of as a conduit in gene expression, in ciliates, RNA can provide both an organizing role in DNA rearrangements and a template for transmitting mutations to the next generation.  Understanding how information is encoded to reorder genome fragments highlights the deep connections between computational and genomic processes.  I explain how a transiently-expressed cache of non-coding RNAs may provide the programming instructions for genome remodeling and transmit heritable information.  The occasional transfer of point mutations from these RNA templates to rearranged DNA molecules supplies a viable mechanism for stable inheritance of acquired characters. This mechanism for inheritance beyond the conventional DNA genome can pass information across multiple generations, hinting at the power of RNA molecules to reshape genome information.

    Related Posts

      A grid-enabled asynchronous metamodel-assisted evolutionary algorithm for aerodynamic optimization

      Abstract  A Grid-enabled asynchronous metamodel-assisted evolutionary algorithm is presented and assessed on a number of aerodynamic
      shape optimization problems. An efficient way of implementing surrogate evaluation models or metamodels (art…

      Abstract  A Grid-enabled asynchronous metamodel-assisted evolutionary algorithm is presented and assessed on a number of aerodynamic
      shape optimization problems. An efficient way of implementing surrogate evaluation models or metamodels (artificial neural
      networks) in the context of an asynchronous evolutionary algorithm is proposed. The use of metamodels relies on the inexact
      pre-evaluation technique already successfully applied to synchronous (i.e. generation-based) evolutionary algorithms, which
      needs to be revisited so as to efficiently cooperate with the asynchronous search method. The so-created asynchronous metamodel-assisted
      evolutionary algorithm is further enabled for Grid Computing. The Grid deployment of the algorithm relies on three middleware
      layers: GridWay, Globus Toolkit and Condor. Single- and multi-objective CFD-based designs of isolated airfoils and compressor cascades are handled using the proposed algorithm and the gain in CPU cost is demonstrated.

      • Content Type Journal Article
      • Pages 373-389
      • DOI 10.1007/s10710-009-9090-5
      • Authors
        • V. G. Asouti, National Technical University of Athens Parallel CFD & Optimization Unit, Lab. of Thermal Turbomachines, School of Mechanical Engineering P.O. Box 64069 Athens 15710 Greece
        • I. C. Kampolis, National Technical University of Athens Parallel CFD & Optimization Unit, Lab. of Thermal Turbomachines, School of Mechanical Engineering P.O. Box 64069 Athens 15710 Greece
        • K. C. Giannakoglou, National Technical University of Athens Parallel CFD & Optimization Unit, Lab. of Thermal Turbomachines, School of Mechanical Engineering P.O. Box 64069 Athens 15710 Greece