Neurocomputing for Design Automation

Angebote / Angebote:

Neurocomputing for Design Automation outlines innovative design theories and computational models with two broad objectives: automation and optimization. This singular book: o Presents an introduction to the automation and optimization of engineering design of complex engineering systems using neural network computing o Outlines new computational models and paradigms for automating the complex process of design for unique engineering systems, such as steel highrise building structures o Applies design theories and models to the solution of structural design problems o Integrates three computing paradigms: mathematical optimization, neural network computing, and parallel processing. The applications described are general enough to be applied directly or by extension to other engineering design problems, such as aerospace or mechanical design. Also, the computational models are shown to be stable and robust - particularly suitable for design automation of large systems, such as a 144-story steel super-highrise building structure with more than 20, 000 members. The book provides an exceptional framework for the automation and optimization of engineering design, focusing on a new computing paradigm - neural networks computing. It presents the automation of complex systems at a new and higher level never achieved before. TOC:Introduction.- Perception.- An Improved Perception Learning Algorithm.- Counter Propagation Neural Network and Its Application to Structural Analysis and Design.- Neural Dynamics Model for Design Optimization - A Design Theory.- Plastic Design of Frames by Neural Dynamics.- Hybrid CPN-Neural Dynamics Model for Discrete Optimization of Steel Structures.- Data Parallel Neural Dynamics Model for Design Optimization on a Massively Parallel Machine.- Distributed Neural Dynamics Model for Design Automation of Large Systems.- References.- Appendix - Sample Codes.