Data-Driven Heuristic Induction From Human Design Behavior

2020 ◽  
Vol 21 (2) ◽  
Author(s):  
Lucas Puentes ◽  
Jonathan Cagan ◽  
Christopher McComb
Keyword(s):  
Cooling System ◽  
Design Problems ◽  
Future Design ◽  
Design Heuristics ◽  
Design Changes ◽  
Induction Process ◽  
Induction Technique ◽  
Applied Design ◽  
Multiple Domains ◽  
Aid Decision

Abstract Through experience, designers develop guiding principles, or heuristics, to aid decision-making in familiar design domains. Generalized versions of common design heuristics have been identified across multiple domains and applied by novices to design problems. Previous work leveraged a sample of these common heuristics to assist in an agent-based design process, which typically lacks heuristics. These predefined heuristics were translated into sequences of specifically applied design changes that followed the theme of the heuristic. To overcome the upfront burden, need for human interpretation, and lack of generality of this manual process, this paper presents a methodology that induces frequent heuristic sequences from an existing timeseries design change dataset. Individual induced sequences are then algorithmically grouped based on similarity to form groups that each represent a shared general heuristic. The heuristic induction methodology is applied to data from two human design studies in different design domains. The first dataset, collected from a truss design task, finds a highly similar set of general heuristics used by human designers to that which was hand-selected for the previous computational agent study. The second dataset, collected from a cooling system design problem, demonstrates further applicability and generality of the heuristic induction process. Through this heuristic induction technique, designers working in a specified domain can learn from others’ prior problem-solving strategies and use these strategies in their own future design problems.

Automated Heuristic Induction From Human Design Data

2020 ◽  
Author(s):  
Lucas Puentes ◽  
Jonathan Cagan ◽  
Christopher McComb
Keyword(s):  
Cooling System ◽  
Design Data ◽  
Design Problems ◽  
Future Design ◽  
Design Changes ◽  
Induction Process ◽  
Induction Technique ◽  
Applied Design ◽  
Multiple Domains ◽  
Aid Decision

Abstract Through experience, designers develop guiding principles, or heuristics, to aid decision-making in familiar design domains. Generalized versions of common design heuristics have been identified across multiple domains and applied by novices to design problems. Previous work leveraged a sample of these common heuristics to assist in an agent-based design process, which typically lacks heuristics. These predefined heuristics were translated into sequences of specifically applied design changes that followed the theme of the heuristic. To overcome the upfront burden, need for human interpretation, and lack of generality of this manual process, this paper presents a methodology that induces frequent heuristic sequences from an existing timeseries design change dataset. Individual induced sequences are then algorithmically grouped based on similarity to form groups that each represent a shared general heuristic. The heuristic induction methodology is applied to data from two human design studies in different design domains. The first dataset, collected from a truss design task, finds a highly similar set of general heuristics used by human designers to that which was hand selected for the previous computational agent study. The second dataset, collected from a cooling system design problem, demonstrates further applicability and generality of the heuristic induction process. Through this heuristic induction technique, designers working in a specified domain can learn from others’ prior problem-solving strategies and use these strategies in their own future design problems.

scholarly journals Design Principles: Literature Review, Analysis, and Future Directions

2016 ◽  
Vol 138 (10) ◽  
Author(s):  
Katherine K. Fu ◽  
Maria C. Yang ◽  
Kristin L. Wood
Keyword(s):  
Design Science ◽  
Design Principles ◽  
Grand Challenge ◽  
Formal Approach ◽  
Design Problems ◽  
Future Directions ◽  
Future Design ◽  
The Future ◽  
Tools And Techniques ◽  
Archival Practices

Design principles are created to codify and formalize design knowledge so that innovative, archival practices may be communicated and used to advance design science and solve future design problems, especially the pinnacle, wicked, and grand-challenge problems that face the world and cross-cutting markets. Principles are part of a family of knowledge explication, which also include guidelines, heuristics, rules of thumb, and strategic constructs. Definitions of a range of explications are explored from a number of seminal papers. Based on this analysis, the authors pose formalized definitions for the three most prevalent terms in the literature—principles, guidelines, and heuristics—and draw more definitive distinctions between the terms. Current research methods and practices with design principles are categorized and characterized. We further explore research methodologies, validation approaches, semantic principle composition through computational analysis, and a proposed formal approach to articulating principles. In analyzing the methodology for discovering, deriving, formulating, and validating design principles, the goal is to understand and advance the theoretical basis of design, the foundations of new tools and techniques, and the complex systems of the future. Suggestions for the future of design principles research methodology for added rigor and repeatability are proposed.

scholarly journals Application of Anti-Stigma Design Heuristics for Usability Inspection

2021 ◽  
Author(s):  
Abdul-Fatawu Abdulai ◽  
A Fuchsia Howard ◽  
Heather Noga ◽  
Paul J Yong ◽  
Leanne M Currie
Keyword(s):  
Health Care ◽  
User Interface ◽  
Health Conditions ◽  
Design Team ◽  
End User ◽  
Design Heuristics ◽  
Usability Problems ◽  
Sexual Pain ◽  
Design Changes ◽  
User Interface Evaluation

User interface evaluation has become important in developing usable health care technologies. Although usability engineering methods have been applied in the design and evaluation of health care software, available heuristics focus on task-work aspects and do not address stigma associated with many health conditions. We used a previous set of heuristics and propose a new set of anti-stigma heuristics to evaluate stigmatization in health care websites. The extended set of heuristics were concurrently applied in a heuristic evaluation and a cognitive walkthrough to evaluate an endometriosis and sexual pain website. The walkthrough involved 5 tasks that required 21 actions to execute. Twenty-six usability problems were identified and recommendations for re-design were made to the design team before end-user testing. The anti-stigma heuristics received worse ratings than the traditional heuristics, resulting in several design changes that might otherwise have been missed. Thus, the new anti-stigma heuristics were a valuable contribution.

Quantity Dimension Indexing for Design Knowledge Management

2007 ◽  
Author(s):  
Tamotsu Murakami ◽  
Yasushi Suehisa
Keyword(s):  
Knowledge Management ◽  
Dimensional Space ◽  
International System ◽  
Design Knowledge ◽  
Design Problems ◽  
Future Design ◽  
Similarity Estimation ◽  
System Of Units ◽  
Physical Phenomena ◽  
Physical Quantities

Although many knowledge management techniques based on text expression have been developed, they are not necessarily sufficient for managing engineering design knowledge. In this paper, we propose quantity dimension indexing of design knowledge as a fundamental method for design knowledge management. Physical quantities describing physical phenomena can be represented as vectors in a seven-dimensional space where the orthogonal axes are the seven base units of the SI (The International System of Units). Because of the generality, objectivity and universality of the SI, this space covers all physical quantities that appear in the past, present and future design knowledge and design problems, and the same quantities are represented as the same vectors regardless of the differences in people, products, domains, organizations, nations and languages. We assume that the similarities of physical phenomena lead to similarities in the dimensions of quantities describing the phenomena, and propose to use this seven-dimensional vector for estimating the similarity of design knowledge from the viewpoint of physical phenomena. Based on this basic idea, we mathematically define similarity between two quantities using quantity dimensions. We prepared design knowledge examples and retrieval keys and conducted design knowledge retrieval and design knowledge similarity estimation by quantity dimension indexing and confirmed that we obtained adequate results without using a concept dictionary or thesaurus elaborated in advance, which are indispensable in the text approach.

FIDOE: A Framework for Intelligent Distributed Ontologies in Engineering

2008 ◽  
Author(s):  
Paul Witherell ◽  
Sundar Krishnamurty ◽  
Ian R. Grosse ◽  
Jack Wileden
Keyword(s):  
Printed Circuit Board ◽  
Circuit Board ◽  
Distributed Environment ◽  
Design Problems ◽  
Design Changes ◽  
Pcb Design ◽  
Common Domain ◽  
Distributed Process ◽  
Solid Foundation ◽  
Future Work

This paper presents FIDOE, a Framework for Intelligent Distributed Ontologies in Engineering. FIDOE consists of a suite of logic rules and templates for interactively developing relationships between properties of linked ontologies. The logical rules embedded in FIDOE automatically operate on various discipline-specific ontologies to systematically identify influences, direct and indirect, of proposed design modifications on other aspects of the design through common domain concepts. Once potential influences are identified, FIDOE enables the user to precisely define the domain relationships, using predefined templates, between the identified domain concepts that enumerate influence types. This tool, thus, provides a pervasive, real time awareness of the implications of design changes during the design process in a distributed environment. The application of FIDOE to distributed and multidisciplinary design problems is detailed with the aid of an industry-provided printed circuit board (PCB) design. Here, commonalities among indirectly connected domain ontologies (electrical, mechanical and thermal domains) are identified using the developed query method and subsequent relationships are defined. These relationships are then applied to provide a collaborative understanding and awareness of the distributed process, all while demonstrating the effectiveness of this approach. This awareness was successfully able to address some previously identified industry concerns, returning promising results while laying a solid foundation for future work.

Towards Orthogonal Turbomachinery Design

2006 ◽  
Author(s):  
Jerome P. Jarrett ◽  
Theo A. Bell ◽  
P. John Clarkson
Keyword(s):  
Design Optimization ◽  
Orthogonal Design ◽  
Pressure Rise ◽  
Global Optimum ◽  
Aviation Industry ◽  
Design Problems ◽  
Surge Margin ◽  
Design Changes ◽  
Engine Compressor ◽  
Turbomachinery Design

The aviation industry is under significant commercial and environmental pressure to produce a revolution in design. However, despite the significant advances in automatic design optimization made over the last 30 years, the industry is still largely conducting design by evolution. The complexity of a modern aeroengine encourages the separation of its conceptual from its detailed design: this limits the utility of powerful design optimization tools solving the “classical” optimization problem (of design space search for the global optimum) to the detailed designer who is more usually tasked with reaching a specification. One of the principal difficulties of modifying the design to reach a particular specified goal is that, though the desired improvement might be achieved, it often comes at unacceptable detriment to other performance indicators. We present results of our orthogonal design technique (that assists the designer in producing improvements in specific attributes of the design without penalty in other aspects) applied to the redesign of a generic core engine compressor for two “real-world” design problems: reducing the part count without aerodynamic penalty and increasing the efficiency without reduction in surge margin, pressure rise or mass flow. The two resulting designs, while meeting their constraints, exhibit a reduction in blading equivalent to two rotor rows and an increase in adiabatic efficiency of 1.0 percentage point respectively. The design changes which produce these improvements, together with how these compare with design rationale, are discussed.

scholarly journals Commercial Aircraft Electrification—Current State and Future Scope

Energies ◽  
2021 ◽  
Vol 14 (24) ◽  
pp. 8381
Author(s):  
Liya Tom ◽  
Muhammad Khowja ◽  
Gaurang Vakil ◽  
Chris Gerada
Keyword(s):  
Fault Tolerant ◽  
Cooling System ◽  
High Strength ◽  
Hydraulic Systems ◽  
Commercial Aircraft ◽  
Future Design ◽  
Current State ◽  
Electric Aircraft ◽  
Increasing Demand ◽  
Drive Systems

Electric and hybrid-electric aircraft propulsion are rapidly revolutionising mobility technologies. Air travel has become a major focus point with respect to reducing greenhouse gas emissions. The electrification of aircraft components can bring several benefits such as reduced mass, environmental impact, fuel consumption, increased reliability and quicker failure resolution. Propulsion, actuation and power generation are the three key areas of focus in more electric aircraft technologies, due to the increasing demand for power-dense, efficient and fault-tolerant flight components. The necessity of having environmentally friendly aircraft systems has promoted the aerospace industry to use electrically powered drive systems, rather than the conventional mechanical, pneumatic or hydraulic systems. In this context, this paper reviews the current state of art and future advances in more electric technologies, in conjunction with a number of industrially relevant discussions. In this study, a permanent magnet motor was identified as the most efficient machine for aircraft subsystems. It is found to be 78% and 60% more power dense than switch-reluctant and induction machines. Several development methods to close the gap between existing and future design were also analysed, including the embedded cooling system, high-thermal-conductivity insulation materials, thin-gauge and high-strength electrical steel and integrated motor drive topology.

scholarly journals The influence of efficiency of the cooling system on the thermodynamic parameters and performance of a two - stage VC 20.96 reciprocating compressor designed to serve as a marine engine starter

2013 ◽  
Vol 20 (4) ◽  
pp. 25-33
Author(s):  
Mateusz Grzelczak
Keyword(s):  
Cooling System ◽  
Pressure Ratio ◽  
Reciprocating Compressor ◽  
Two Stage ◽  
Ratio Distribution ◽  
Design Work ◽  
Marine Engine ◽  
Type Series ◽  
Basic Task ◽  
Applied Design

ABSTRACT The paper presents results of the research related to the analysis of the thermodynamic and flow processes occurring in a prototype VC 20.96 two-stage liquid-cooled reciprocating compressor. The compressor has been developed and manufactured by H. Cegielski Poznan metal works in collaboration with the Poznan University of Technology. The research related to the VC compressor was realized within the KBN 3127/C.T07-6/2002 project titled “Development of design of type-series of reciprocating compressors and their implementation in production”. The basic task of the project was to develop two type-series of liquid- and air- cooled reciprocating compressors of the V- and W- arrangement, designed to serve as marine engine starters. The result of the design work was the manufacturing of two compressors: the VC 20.96 liquid-cooled compressor and the WP 18.80 air-cooled one. The main aim of the research described in this paper was to evaluate the efficiency of the cooling system which uses inter-coolers integrated with the compressing stages and the cooperation of the compressing stages in terms of pressure ratio distribution. Owing to the cooling method, the applied design assumptions enabled to develop a compact compressor fulfilling the assumed operating parameters.

scholarly journals Improving De Novo Molecular Design with Curriculum Learning

2021 ◽  
Author(s):  
Jeff Guo ◽  
Vendy Fialková ◽  
Juan Diego Arango ◽  
Christian Margreitter ◽  
Jon Paul Janet ◽  
...  
Keyword(s):  
De Novo ◽  
Molecular Design ◽  
Positive Impact ◽  
Accelerated Learning ◽  
Design Problems ◽  
Suitable Alternative ◽  
De Novo Molecular Design ◽  
Multiple Domains ◽  
Multiple Interactions

Abstract Reinforcement learning (RL) is a powerful paradigm that has gained popularity across multiple domains. However, applying RL may come at a cost of multiple interactions between the agent and the environment. This cost can be especially pronounced when the single feedback from the environment is slow or computationally expensive, causing extensive periods of nonproductivity. Curriculum learning (CL) provides a suitable alternative by arranging a sequence of tasks of increasing complexity with the aim of reducing the overall cost of learning. Here, we demonstrate the application of CL for drug discovery. We implement CL in the de novo design platform, REINVENT, and apply it on illustrative de novo molecular design problems of different complexity. The results show both accelerated learning and a positive impact on the quality of the output when compared to standard policy based RL. To our knowledge, this is the first application of CL for the purposes of de novo molecular design. The code is freely available at https://github.com/MolecularAI/Reinvent.

Extending Design Heuristics From Mechanical Engineering to a Biomedical Projects Course

2017 ◽  
Author(s):  
Jin Woo Lee ◽  
Anastasia K. Ostrowski ◽  
Shanna R. Daly ◽  
Aileen Y. Huang-Saad ◽  
Colleen M. Seifert
Keyword(s):  
Engineering Design ◽  
Biomedical Engineering ◽  
Empirical Studies ◽  
Idea Generation ◽  
Instructional Tools ◽  
Design Heuristics ◽  
Award Winning ◽  
Individual Concepts ◽  
Applied Design ◽  
Support Students

Research in design has led to emergence of instructional tools to support students in generating multiple candidate concepts. Design Heuristics was developed through empirical studies of professional engineers and award-winning products, and have been shown to support student engineers in generating creative and diverse concepts. We hypothesized that they could be beneficial to student designers in biomedical engineering. In this qualitative classroom study, we examined how graduate students in a biomedical engineering design course applied Design Heuristics to generate individual concepts for their design projects. Our analysis showed that students were able to apply Design Heuristics in their biomedical engineering projects, and that the heuristics supported idea generation in a variety of biomedical engineering design contexts.

Sign in / Sign up

Export Citation Format

Share Document