Title :
Self-organizing nets for optimization
Author :
Milano, Michele ; Koumoutsakos, Petros ; Schmidhuber, Jürgen
Author_Institution :
Graduate Aeronaut. Labs., California Inst. of Technol., Pasadena, CA, USA
fDate :
5/1/2004 12:00:00 AM
Abstract :
Given some optimization problem and a series of typically expensive trials of solution candidates sampled from a search space, how can we efficiently select the next candidate? We address this fundamental problem by embedding simple optimization strategies in learning algorithms inspired by Kohonen´s self-organizing maps and neural gas networks. Our adaptive nets or grids are used to identify and exploit search space regions that maximize the probability of generating points closer to the optima. Net nodes are attracted by candidates that lead to improved evaluations, thus, quickly biasing the active data selection process toward promising regions, without loss of ability to escape from local optima. On standard benchmark functions, our techniques perform more reliably than the widely used covariance matrix adaptation evolution strategy. The proposed algorithm is also applied to the problem of drag reduction in a flow past an actively controlled circular cylinder, leading to unprecedented drag reduction.
Keywords :
covariance matrices; learning (artificial intelligence); optimisation; self-organising feature maps; stochastic processes; Kohonen self-organizing maps; active data selection process; adaptation evolution strategy; adaptive grids; adaptive nets; circular cylinder control; covariance matrix; drag reduction; net nodes; neural gas networks; optimization problem; search space regions; self-organizing nets; Adaptive systems; Convergence; Covariance matrix; Genetic algorithms; Genetic mutations; Learning; Mesh generation; Optimization methods; Self organizing feature maps; Stochastic processes; Neural Networks (Computer);
Journal_Title :
Neural Networks, IEEE Transactions on
DOI :
10.1109/TNN.2004.826132