A Machine Learning-based Tool for Generation and Evaluation of Product Design Concepts

The global competitive market requires products to meet the high customer satisfaction, low production cost and short development cycle. Product design searches for potential solutions of products to meet customer requirements based on design constraints and technical details. As a large number of possible concepts have to be explored in the design process, applying the existing approaches and tools is a costly and time-consuming process. To overcome the limitations, a novel approach is proposed based on reinforcement learning to train intelligent agents for learning design. The approach will apply the existing design knowledge and available resources for product functions to generate the optimal design solution for reduced cost and time in the product design. A case study of the hand rehabilitation device design will be used to test the accuracy and effectiveness of the proposed tool.

Faculty Supervisor:

Qingjin Peng

Student:

Partner:

North Forge

Discipline:

Engineering

Sector:

Education; Management of companies and enterprises; Professional, scientific and technical services

University:

University of Manitoba

Program: