We developed an efficient semisupervised feedforward neural network clustering model with one epoch training and data dimensionality reduction ability to solve the problems of low training speed, accuracy, and high memory complexity of clustering. During training, a codebook of nonrandom weights is learned through input data directly. A standard weight vector is extracted from the codebook, and the exclusive threshold of each input instance is calculated based on the standard weight vector. The input instances are clustered based on their exclusive thresholds. The model assigns a class label to each input instance through the training set. The class label of each unlabeled input instance is predicted by considering a linear activation function and the exclusive threshold. Finally, the number of clusters and the density of each cluster are updated. The accuracy of the proposed model was measured through the number of clusters and the quantity of correctly classified nodes, which was 99.85%, 100%, and 99.91% of the Breast Cancer, Iris, and Spam data sets from the University of California at Irvine Machine Learning Repository, respectively, and the superior F measure results between 98.29% and 100% accuracies for the breast cancer data set from the University of Malaya Medical Center to predict the survival time.

The mechanism of design reasoning from function to form is suggested to consist of a two-step inference of the innovative abduction type. First is an inference from a desired functional aspect to an idea, concept, or solution principle to satisfy the function. This is followed by a second innovative abduction, from the latest concept to form, structure, or mechanism. The intermediate entity in the logical reasoning, the concept, is thus made explicit, which is significant in following and understanding a specific design process, for educating designers, and to build a logic-based computational model of design. The idea of a two-step abductive reasoning process is developed from the critical examination of several propositions made by others. We use the notion of innovative abduction in design, as opposed to such abduction where the question is about selecting among known alternatives, and we adopt a previously proposed two-step process of abductive reasoning. However, our model is different in that the two abductions used follow the syllogistic pattern of innovative abduction. In addition to using a schematic example from the literature to demonstrate our derivation, we apply the model to an existing, empirically derived method of conceptual design called “parameter analysis” and use two examples of real design processes. The two synthetic steps of the method are shown to follow the proposed double innovative abduction scheme, and the design processes are presented as sequences of double abductions from function to concept and from concept to form, with a subsequent deductive evaluation step.

Many of the most fundamental decisions about a product are made during the system architecture design process. However, how system architecture is designed in practice is not well understood. This paper draws on several research studies related to system architecture design to develop a categorization of system architecture design processes to support the adaptation design methodologies and tools to specific situations. The paper reviews different definitions of system architecture and comments on the relevance of the different perspectives taken in the literature on system architecture to different types of system architecture. The research highlights the need for further empirical research on system architecture design processes as well as on tools to support the engineers creating the system architecture.

Stylistic composition is a creative musical activity, in which students as well as renowned composers write according to the style of another composer or period. We describe and evaluate two computational models of stylistic composition, called Racchman-Oct2010 (random constrained chain of Markovian nodes, October 2010) and Racchmaninof-Oct2010 (Racchman with inheritance of form). The former is a constrained Markov model, and the latter embeds this model in an analogy-based design system. Racchmaninof-Oct2010 applies a pattern discovery algorithm called SIACT and a perceptually validated formula for rating pattern importance, to guide the generation of a new target design from an existing source design. A listening study is reported concerning human judgments of music excerpts that are, to varying degrees, in the style of mazurkas by Frédéric Chopin (1810–1849). The listening study acts as an evaluation of the two computational models and a third, benchmark system, called Experiments in Musical Intelligence. Judges' responses indicate that some aspects of musical style, such as phrasing and rhythm, are being modeled effectively by our algorithms. Judgments are also used to identify areas for future improvements. We discuss the broader implications of this work for the fields of engineering and design, where there is potential to make use of our models of hierarchical repetitive structure.

Requirement planning is one of the most critical tasks in the product development process. Despite its significant impact on the outcomes of the design process, engineering requirement planning is often conducted in an ad hoc manner without much structure. In particular, the requirement planning phase suffers from a lack of quantifiable measures for evaluating the quality of the generated requirements and also a lack of structure and formality in representing engineering requirements. The main objective of this research is to develop a formal Web Ontology Language ontology for standard representation of engineering requirements. The proposed ontology uses explicit semantics that makes the ontology amenable to automated reasoning. To demonstrate how the proposed ontology can support requirement analysis and evaluation in engineering design, three possible services enabled by the ontology are introduced in this paper. These services are information content measurement, specificity and completeness analysis, and requirement classification. The proposed ontology and its associated algorithms and tools are validated experimentally in this work.

Prior research has identified set-based design as a method that accounts for the high level of uncertainty that is associated with the design of innovative products or systems. Rather than precisely specifying a system architecture in the early design stages, set-based design builds on designing a system and its architecture in an evolutionary way. The literature on set-based design has studied how a system's design evolves by moving from a number of optional design ideas to the final system through gradually eliminating unfeasible design ideas and continually developing design ideas for which engineers increasingly establish feasibility. However, little is known about how firms set up the design process and the organization to successfully create new products with set-based design. Our research contributes to closing this gap. First, we study how firms determine the number (i.e., portfolio) of design ideas to pursue, an important step of the early design process. Second, we study how firms organize for set-based design by assigning teams to develop design ideas and eventually design a system's architecture. Our research uses an exploratory case study approach, investigating five cases in three different firms. First, we find that the early design process is characterized by the absence of formal idea evaluation and selection. Instead, firms start to pursue all initially created design ideas, evaluating and selecting them in an evolutionary manner as the design project progresses. Second, we identify two organizational approaches associated with set-based design: assign one team to pursue all ideas or assign one team per design idea.

One of the biggest challenges associated with design and digital reconstruction of free forms comes from uniqueness and unrepeatability of these shapes. During digital reconstruction of these forms, the designer has to choose the right set of geometric features and then compose them in a way that will enable the most accurate reconstruction of the geometry. While doing this, the designer primarily relies on personal experience gained through work with free-form objects of similar geometry. In our opinion, the analysis of free-form objects geometry should rely upon semantic interpretation of their geometric and other features, and the greatest challenge of automation of digital reconstruction and free-form object design in general is closely related to automation of semantic interpretation of geometric and other free-form object features. In this paper, a case of chest bone implant digital reconstruction is presented, where a new semantic model called the active semantic model was used for modeling the meaning of geometric elements, that is, the semantic features of a free-form object. The active semantic model and its analogy-based reasoning algorithms have shown themselves as applicable for the automation of semantic interpretation of the unique, unrepeatable, and unpredictable forms of chest bone. Moreover, this semantic model showed the potential to help automate selecting and composing of geometric features for efficient digital reconstruction of the geometry of free forms.

Shape grammars are rule-based formalisms for the specification of shape languages. Most of the existing shape grammars are developed on paper and have not been implemented computationally thus far. Nevertheless, the computer implementation of shape grammar is an important research question, not only to automate design analysis and generation, but also to extend the impact of shape grammars toward design practice and computer-aided design tools. In this paper, we investigate the implementation of shape grammars on a computer system, using a graph-theoretic representation. In particular, we describe and evaluate the implementation of the existing Rabo-de-Bacalhau transformation grammar. A practical step-by-step approach is presented, together with a discussion of important findings noticed during the implementation and evaluation. The proposed approach is shown to be both feasible and valuable in several aspects: we show how the attempt to implement a grammar on a computer system leads to a deeper understanding of that grammar, and might result in the further development of the grammar; we show how the proposed approach is embedded within a commercial computer-aided design environment to make the shape grammar formalism more accessible to students and practitioners, thereby increasing the impact of grammars on design practice; and the proposed step-by-step implementation approach has shown to be feasible for the implementation of the Rabo-de-Bacalhau transformation grammar, but can also be generalized using different ontologies for the implementation.

Is biologically inspired design only an analogical transfer from biology to engineering? Actually, nature does not always bring “hands-on” solutions that can be analogically applied in classic engineering. Then, what are the different operations that are involved in the bioinspiration process and what are the conditions allowing this process to produce a bioinspired design? In this paper, we model the whole design process in which bioinspiration is only one element. To build this model, we use a general design theory, concept–knowledge theory, because it allows one to capture analogy as well as all other knowledge changes that lead to the design of a bioinspired solution. We ground this model on well-described examples of biologically inspired designs available in the scientific literature. These examples include Flectofin®, a hingeless flapping mechanism conceived for façade shading, and WhalePower technology, the introduction of bumps on the leading edge of airfoils to improve aerodynamic properties. Our modeling disentangles the analogical aspects of the biologically inspired design process, and highlights the expansions occurring in both knowledge bases, scientific (nonbiological) and biological, as well as the impact of these expansions in the generation of new concepts (concept partitioning). This model also shows that bioinspired design requires a special form of collaboration between engineers and biologists. Contrasting with the classic one-way transfer between biology and engineering that is assumed in the literature, the concept–knowledge framework shows that these collaborations must be “mutually inspirational” because both biological and engineering knowledge expansions are needed to reach a novel solution.

Although there has been considerable computer-aided conceptual design research, most of the proposed approaches are domain specific and can merely achieve conceptual design of energy flows-processing systems. Therefore, this research is devoted to the development of a general (i.e., domain-independent) and knowledge-based methodology that can search in a wide multidisciplinary solution space for suitable solution principles for desired material-flow processing functions without designers' biases toward familiar solution principles. It first proposes an ontology-based approach for representing desired material-flow processing functions in a formal and unambiguous manner. Then a rule-based approach is proposed to represent the functional knowledge of a known solution principle in a general and flexible manner. Thereafter, a simulation-based retrieval approach is developed, which can search for suitable solution principles for desired material-flow processing functions. The proposed approaches have been implemented as a computer-aided conceptual design system for test. The conceptual design of a coin-sorting device demonstrates that our functional representation methodology can make the proposed computer-aided conceptual design system to effectively and precisely retrieve suitable solution principles for a desired material-flow processing function.

Consumers' psychological perceptions of a product are significantly influenced by its appearance aesthetics, and thus product form plays an essential role in determining the commercial success of a product. The evolution of a product's form during the design process is typically governed by the designer's individual preferences and creative instincts. As a consequence, there is a risk that the product form may fail to satisfy the consumers' expectations or may induce an unanticipated consumer response. This study commences developing an integrated design approach based on the numerical definition of product form. A series of evaluation trials are then performed to establish the correlation between the product form features and the consumers' perceptions of the product image. The results of the evaluation trials are used to construct three different types of mathematical model (a multiple regression analysis model, a backpropagation neural network model, and a multiple regression analysis with a backpropagation neural network model) to predict the likely consumer response to any arbitrary product form. The feasibility of an integrated design approach is demonstrated using a three-dimensional knife form. Although this study takes an example for illustration and verification purposes, the methodology proposed in the present study is equally applicable to any form of consumer product.

Decisions related to system architecture are difficult because of fuzziness and lack of information combined with often-conflicting objectives. We organized an industrial workshop with the objective of choosing 5 out of 800 architectures. The first step, the identification of selection criteria, proved to be the greatest challenge. As a result, designers selected system architectures that did not satisfy them without being able to explain why. It appeared that most of the difficulties faced by the designers came from the criteria used for architecture selection. This study aims to identify what made the selection criteria difficult to use. The audio recordings of the workshop were transcribed and analyzed in order to identify the obstacles related to the definition and the use of selection criteria. The analysis highlights two issues: the interdisciplinarity of system architecture makes criteria interdependent and the lack of information makes it impossible to define an exhaustive set of criteria. Finally, this study provides recommendations for selecting appropriate selection criteria and insights for future selection support tools dedicated to system architecture design.

Design fixation is a factor that negatively influences the generation of novel design concepts (Jansson & Smith, 1991). When designers fixate, they tend to reproduce example features or features from their initial ideas. In order to mitigate design fixation, it is crucial to identify the factors that influence the extent of design fixation. This paper investigates two such factors: the modality of examples and the familiarity of designers with the example features. To investigate this, an experiment is conducted with mechanical engineering students who were asked to generate ideas to solve a peanut sheller design problem. The students generated ideas in five different experimental conditions: control, where no example was given; the first example given in a sketch form; the first example given as a nonfunctional prototype; a second example in sketch form; and the second example in a working prototype form. The first example was a nonfeasible solution, but it contained several features familiar to the participants. The second example was a feasible solution, but it contained less familiar features. In order to understand the extent of fixation triggered by the examples, three metrics were utilized to compare across the experimental conditions: the quantity of nonredundant ideas generated by the participants, the presence of example features in their solutions, and their fixation to the example's energy source. The results showed that in the case of the familiar example, the example modality did play an important role in the extent of design fixation. Across the examples, it was found that the first example containing several familiar features caused more fixation than the second one. Overall, this paper shows that the modality in which the example was communicated and the presence of familiar features in an example influenced the fixation caused by those examples.

Maintenance plays a critical role in reducing operating cost and maximizing reliability of a complex engineering system. This paper proposes a novel maintenance-focused, system-level design framework that attempts to capture the interactions between maintenance strategies and system-level design parameters overlooked in current modeling approaches. The goal of this maintenance-focused approach is to help designers better understand the interconnectedness of system architecture, choice of maintenance strategy, and uncertainties in a design. Application of the proposed design framework is demonstrated through a case example of a power plant condenser system. Results show that using an integrated approach can reveal the many nonobvious interactions between subsystems, and produce system designs that have lower life-cycle cost compared to traditional sequential design approaches.

This paper presents a method for automatically generating new designs from a set of existing objects of the same class using machine learning. In this particular work, we use a custom parametric chair design program to produce a large set of chairs that are tested for their physical properties using ergonomic simulations. Design schemata are found from this set of chairs and used to generate new designs by placing constraints on the generating parameters used in the program. The schemata are found by training decision trees on the chair data sets. These are automatically reverse engineered by examining the structure of the trees and creating a schema for each positive leaf. By finding a range of schemata, rather than a single solution, we maintain a diverse design space. This paper also describes how schemata for different properties can be combined to generate new designs that possess all properties required in a design brief. The method is shown to consistently produce viable designs, covering a large range of our design space, and demonstrates a significant time saving over generate and test using the same program and simulations.

As more and more people are increasingly turning to nature for design inspiration, tools and methodologies are developed to support the systematic bioideation process. State-of-the-art approaches struggle with expanding their knowledge bases because of interactive work required by humans per biological strategy. As an answer to this persistent challenge, a scalable search for systematic biologically inspired design (SEABIRD) system is proposed. This system leverages experience from the product aspects in design by analogy tool that identifies candidate products for between-domain design by analogy. SEABIRD is based on two conceptual representations, product and organism aspects, extracted from, respectively, a patent and a biological database, that enable leveraging the ever growing body of natural-language biological texts in the systematic bioinspired design process by eliminating interactive work by humans during corpus expansion. SEABIRD's search is illustrated and validated with three well-known biologically inspired design cases.