A New Constructive Compound Neural Networks Using Fuzzy Logic and Genetic Algorithm 1 Application to Artificial Life

スポンサーリンク

概要

• 論文の詳細を見る

• This paper presents a new compound constructive algorithm of neural networks whereby the fuzzy logic technique is explored as an efficient learning algorithm to implement an optimal network construction from an initial simple 3-layer network while the genetic algorithm is used to help design an improved network by evolutions. Numerical simulations on artificial life demonstrate that compared with the existing network design algorithms such as the constructive algorithms of [1]-[3], the pruning algorithms of [5], [6] and the fixed, static architecture algorithm of [7], the present algorithm, called FuzGa, is efficient in both time complexity and network performance. The improved time complexity comes from the sufficiently small 3 layer design of neural networks and the genetic algorithm adopted partly because the relatively small number of layers facilitates an utilization of an efficient steepest descent method in narrowing down the solution space of fuzzy logic and partly because trappings into local minima can be avoided by genetic algorithm, contributing to considerable saving in time in the processing of network learning and connection. Compared with 54.8 minutes of MLPs with 65 hidden neurons [7], 63.1 minutes of FlexNet [1] or 96.0 minutes of Pruning [5], our simulation results on artificial life show that the CPU time of the present method reaching the target fitness value of 100 food elements eaten for the present FuzGa has improved to 42.3 minutes by SUN's SPARCstation-10 of SuperSPARC 40MHz machine for example. The role of hidden neurons is elucidated in improving the performance level of the neural networks of the various schemes developed for artificial life applications. The effect of population size on the performance level of the present FuzGa is also elucidated.

• 社団法人電子情報通信学会の論文
• 1998-12-25

著者

• TOKUDA Naoyuki

  SunFlare Research and Development Center

• Tokuda N

  Sunflare Research And Development Center

• Tokuda Naoyuki

  Computer Science Department Utsunomiya University Utsunomiya Japan

• Tokuda Naoyuki

  Faculty Of General Education Utsunomiya University

• YAN Jianjun

  The authors are with Computer Science Department, Faculty of Engineering, Utsunomiya University

• YAN Jianjun

  Computer Science Department, Faculty of Engineering, Utsunomiya University

• MIYAMICHI Juichi

  Computer Science Department, Faculty of Engineering, Utsunomiya University

• Miyamichi J

  Faculty Of Engineering Utsunomiya University

• Yan Jianjun

  The Authors Are With Computer Science Department Faculty Of Engineering Utsunomiya University

関連論文

• Instantaneously Reversible Golomb-Rice Codes for Robust Image Coding
• Compound Image Compression Using Adaptive Wavelet Transform
• A Shrinking Method for Dithered Images
• Capacity analysis for a two-level decoupled Hamming network for associative memory under a noisy environment
• Probabilistic Information Retrieval Method Based on Differential Latent Semantic Index Space
• A Constructive Compound Neural Networks. II Application to Artificial Life in a Competitive Environment
• A New Constructive Compound Neural Networks Using Fuzzy Logic and Genetic Algorithm 1 Application to Artificial Life
• A Model-Based Active Landmarks Tracking Method
• Image Morphing by Spatial Thin-Plate Spline Transformation
• A Finite-Difference Solution of a Crystal Growth Stefan Problem
• Stokes Solutions for Flow Near Corners in Three-Dimensions
• A Syntactical Approach to Diagnosing Multiple Bugs in an Intelligent Tutoring System
• A Convection Stefan Problem by Lagrange-Burmann Expansions.I.Small Time Solution

もっと見る 閉じる

スポンサーリンク