Global Topology Optimisation

Abstract

The aim of the proposal is to develop a global topology optimization methodology that can explore global and multiple optima in a design space, within a practical limit of computational resources. All established topology optimization methods employ gradients to search for a single design, which may be either a local or global optimum. Previous methods that attempt to obtain the global optimum have typically utilized non-gradient based population search methods, which have not been successful due to the high dimensional space of topology optimization. This proposal will investigate the feasibility of a global optimization method for multimodal topology optimization inspired by parallel tempering. The objectives are: 1. To develop a global optimization method based on numerical gradients, random noise and parallel tempering. 2. To characterize local and global minima in benchmark multi-modal optimization problems. 3. To investigate the diversity of the local and global minima in topology optimization.

Document Details

Document Type
DoD Grant Award
Publication Date
Mar 23, 2016
Source ID
FA95501510317

Entities

People

  • Robert L. Jack

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • University of Bath

Tags

Readers

  • Operations Research

Technology Areas

  • Space