Abstract:Advancements in prototyping knowledge capture have resulted in several graph-based schemas where nodes represent prototyping activities and individuals, and edges represent semantic relationships between prototypes (e.g. influence and involvement) defined by the design team. Their subsequent implementation in web-based applications has enabled the integration of both physical and virtual prototyping knowledge. The integration affords traceability, version control, and application of graph analysis to generate insights into the Engineering Design process and the schema validation studies reported in prior research have resulted in several open-access Prototyping Knowledge Graph datasets. This paper builds on prior research by contributing a data mining analysis of nine Prototyping Knowledge Graphs. The analysis followed the CRISP-DM process. The purpose was to elicit insights on prototyping behaviour within the design process. The results revealed nine findings. Five corroborated existing observations and four were new observations. In particular, data mining of the knowledge graphs showed sequence and type of knowledge generated from prototyping are unique to each design process. Also, the critical design path is dominated by physical prototypes.

What problem does this paper attempt to address?

This paper primarily explores how to analyze Prototyping Knowledge Graphs (PKGs) in the design process using data mining methods to gain new insights into prototyping behaviors. The following is a summary of the key issues the paper attempts to address: 1. **Research Background and Objectives**: - Prototyping activities generate a large amount of knowledge during the engineering design process, which is crucial for decision-making. - In recent years, the development of prototyping methods and technologies has facilitated the integration of knowledge from both physical and virtual prototypes. - Researchers have developed graph-based knowledge representation methods to capture the knowledge in prototyping activities. These graphs consider prototyping activities and related personnel as nodes and the relationships between prototypes as edges. 2. **Contributions of the Paper**: - This paper analyzes nine Prototyping Knowledge Graphs using data mining, following the CRISP-DM (Cross-Industry Standard Process) methodology. - The aim is to extract insights about prototyping behaviors from the design process. - The analysis revealed nine findings, five of which confirmed existing observations, and four were new observations. 3. **Specific Findings**: - The sequence and type of knowledge generated by prototyping are unique to each design process. - Key design paths are often dominated by physical prototypes. - Physical prototypes are more effective than virtual prototypes in facilitating collaboration. 4. **Methodology**: - Public datasets were used, which came from two four-day design hackathon events. - Graph analysis methods were applied, including centrality measures, critical path analysis, etc., to reveal patterns and structures in the design process. In summary, this paper aims to deeply understand the role of prototyping in the engineering design process through data mining techniques and provide empirical evidence for optimizing the design process.

Data mining prototyping knowledge graphs for design process insights

ProtoDreamer: A Mixed-prototype Tool Combining Physical Model and Generative AI to Support Conceptual Design

A Hybrid Prototype Method Combining Physical Models and Generative Artificial Intelligence to Support Creativity in Conceptual Design

REVISITING PROTOTYPING IN 2020: A SNAPSHOT OF PRACTICE IN UK DESIGN COMPANIES

Why we Prototype! An International Comparison of the Linkage between Embedded Knowledge and Objective Learning

An Artificial Intelligence Based Data-Driven Approach for Design Ideation

Digitally Capturing Physical Prototypes During Early-Stage Engineering Design Projects for Initial Analysis of Project Output and Progression

A Knowledge Graph-Based Approach to Modeling & Representation for Machining Process Design Intent

DesignFusion: Integrating Generative Models for Conceptual Design Enrichment

Developing A Toolkit For Prototyping Machine Learning-Empowered Products: The Design And Evaluation Of Ml-Rapid

People, Projects, Organizations, and Products: Designing a Knowledge Graph to Support Multi-Stakeholder Environmental Planning and Design

Reference model for the development of prototypes based on C-K theory

Prototyping Strategies to Engage Stakeholders During Early Stages of Design: A Study Across Three Design Domains

The Design Exemplar: A New Data Structure for Embodiment Design Automation

Sketch2Prototype: Rapid Conceptual Design Exploration and Prototyping with Generative AI

Content-Centric Prototyping of Generative AI Applications: Emerging Approaches and Challenges in Collaborative Software Teams

Top-Down Hierarchical Construction and Application of a Domain Knowledge Graph based on Multimodal Design Information

A case-based knowledge graph with reinforcement learning for intelligent design approach of complex product

Integration of Knowledge Engineering and 3d Printing for Digital Design and Manufacturing - A Case Study

Jigsaw: Supporting Designers to Prototype Multimodal Applications by Chaining AI Foundation Models

Guiding Data-Driven Design Ideation by Knowledge Distance