This Website is not fully compatible with Internet Explorer.
For a more complete and secure browsing experience please consider using Microsoft Edge, Firefox, or Chrome

Optimization: From Generative to Human-Assisted Design, and Machine Learning

This presentation was made at CAASE18, The Conference on Advancing Analysis & Simulation in Engineering. CAASE18 brought together the leading visionaries, developers, and practitioners of CAE-related technologies in an open forum, to share experiences, discuss relevant trends, discover common themes, and explore future issues.

Resource Abstract

For millenia, up until the last 150 years or so, the only way to evaluate a design was to build and test it. Henry Petroski ("To Engineer is Human") tells it very well, how the British developed iron railroad bridges. Quite simply, they changed the design until the bridges stopped falling down. (https://en.wikipedia.org/wiki/Henry_Petroski)

Then we see the rise of analytical methods that helped with the design, and even allowed evaluation of proposed designs. This is digital prototyping, like we have seen with finite-element analysis for the last sixty years or so.

But now, we have something different. Generative Design is able to take statements of product requirements and transform them into designs (shape, materials, and configurations) in ways that unaided humans cannot. There is conceptually no need to confirm the design, for the design process itself ensures the requirements have been met.

Generative design will be at the nexus of optimization, materials engineering, and new manufacturing methods like additive manufacturing. One also believes that I am being too brief, and I need to add more words here.

Now, the idea is we can add machine learning to the process, to do what some call "Human-Assisted Design". Here, the human is an orchestrator, the conductor, making choices from machine-generated design options. As people like Ted Blacker will eloquently tell you, this will enable innovation in ways we cannot imagine.

To me, this is the important point. We are moving from evaluating proposed designs to directly creating capable designs.

A unified way to think of such methods is Optimization. These methods modify (or even create) a design based on requirements. If one then adds the layer of machine learning or artificial intelligence, I believe we are on the verge of a profound revolution in product design and engineering.

Document Details

ReferenceCAASE_Jun_18_4
AuthorMeintjes. K
LanguageEnglish
TypePresentation
Date 7th June 2018
OrganisationCIMdata Inc
RegionAmericas

Download


Back to Previous Page