Project Dreamcatcher

Nowadays, the conventional way of designing something, from buildings to airplanes, is to have one or multiple designers come up with one or several design options and then pass them on to another group of people for evaluation and feasibility assessments. In essence, this workflow, which is followed in almost all engineering practices, has some limitations.

First of all, after weeks if not months of work the designers may come up with only a handful of designs. This seems very natural due to the limitations of the human brain. Afterward, these designs need to be assessed from different perspectives in other departments to ensure they satisfy the required constraints. For example, a structure must withstand the applied loads to it. Likely, this won’t be the case and the design will be returned to the design department. This back-and-forth process continues several times until a solution complying with all the requirements is finalized.

In project Dreamcatcher, we tried to change this inefficient process. Instead of having someone come up with the design, they define the goals and the constraints that the final solution must satisfy, and it is the computer that iterates through this vast design space, exploiting AI and complex mathematical models, and generates the solution(s) considering all the requirements. This concept is called “generative design“:

Ever since the successful execution of Project Dreamcatcher, generative design has found its way into many industries, revolutionized their usual practices, and pushed its way from a research project to becoming an Autodesk product. Below are some other videos about Autodesk’s generative design (emerging from project Dreamcatcher)!