Title

New concepts in evolutionary algorithms for systems architecture optimization

Authors

Ossama Abdelkhalik, Michigan Technological UniversityFollow

Files

Publication Date

10-5-2017

Description

This talk will present recent developments in evolutionary algorithms that focus on the systems architecture optimization. One challenge in systems optimization is the variable number of design variables. Three new concepts will be presented in this talk that enable evolutionary algorithms to handle this type of problems more efficiently. The first is the biologically inspired concept of Hidden Genes that implements tags to cover/uncover genes in the variables' code, which enable handling variable number of variables in evolutionary algorithms. The second concept is a structured chromosome approach that transcripts the variables in a multi-layer code as opposed to the standard single-layer coding. Finally, a dynamic population size concept is presented. Application of these methods in aerospace engineering will be presented.

Disciplines

Engineering Mechanics | Mechanical Engineering

Recommended Citation

Abdelkhalik, Ossama, "New concepts in evolutionary algorithms for systems architecture optimization" (2017). TechTalks. 51.
https://digitalcommons.mtu.edu/techtalks/51