Complex engineered systems design is a collaborative activity. To design a system, experts from the relevant disciplines must work together to create the best overall system from their individual components. This situation is analogous to a multiagent system in which agents solve individual parts of a larger problem in a coordinated way. Current multiagent models of design teams, however, do not capture this distributed aspect of design teams – instead either representing designers as agents which control all variables, measuring organizational outcomes instead of design outcomes, or representing different aspects of distributed design, such as negotiation. This paper presents a new model which captures the distributed nature of complex systems design by decomposing the ability to control design variables to individual computational designers acting on a problem with shared constraints. These designers are represented as a multiagent learning system which is shown to perform similarly to a centralized optimization algorithm on the same domain. When used as a model, this multiagent system is shown to perform better when the level of designer exploration is not decayed but is instead controlled based on the increase of design knowledge, suggesting that designers in multidisciplinary teams should not simply reduce the scope of design exploration over time, but should adapt based on changes in their collective knowledge of the design space. This multiagent system is further shown to produce better-performing designs when computational designers design collaboratively as opposed to independently, confirming the importance of collaboration in complex systems design.

This study deals with both a decision model for making decisions under epistemic uncertainty and how to use it for selecting optimal materials under the same uncertainty. In particular, the proposed decision model employs a set of possibilistic objective functions defined by fuzzy numbers to handle a set of conflicting criteria. In addition, the model can calculate the compliance of a piece of decision-relevant (imprecise) information with a given objective function. Moreover, the model is capable to aggregate the calculated compliances for the sake of ranking a given set of alternatives against the set of conflicting criteria. The problem of selecting materials for making the body of a vehicle is considered as an example. In this problem, the indices for selecting the materials are unknown because the specifications regarding the vehicle body are not given. In addition, the data relevant to material properties entails a great deal of imprecision. The presented decision model can quantify the above-mentioned epistemic uncertainty in a lucid manner and generate a list of optimal materials.

The research on new types of design knowledge bases and conceptual solution catalogues is an advanced subject of design methodology. Using the separate function as an example, this paper establishes design catalogues for principle conceptual design to realize the function, and proposes a general computer-oriented action principle design process. Upon studying the relationship of knowledge in the process, this paper represents functions of action principles in function feature models and provides computer-oriented feature models for action principle design.

This paper presents a bioinspiration approach that is able to scalably leverage the ever-growing body of biological information in natural-language format. The ideation tool AskNature, developed by the Biomimicry 3.8 Institute, is expanded with an algorithm for automated classification of biological strategies into the Biomimicry Taxonomy, a three-level, hierarchical information structure that organizes AskNature's database. In this way, the manual work entailed by the classification of biological strategies can be alleviated. Thus, the bottleneck is removed that currently prevents the integration of large numbers of biological strategies. To demonstrate the feasibility of building a scalable bioideation system, this paper presents tests that classify biological strategies from AskNature's reference database for those Biomimicry Taxonomy classes that currently hold sufficient reference documents.

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.

This paper discusses the integration of knowledge bases and shape grammars for the generation of building models, covering interaction, system, and implementation. Knowledge-based and generative systems are combined to construct a method for characterizing existing buildings, in particular, their interior layouts based on exterior features and certain other parameters such as location and real dimensions. The knowledge-based model contains information about spatial use, organization, elements, and contextual information, with the shape grammar principally containing style rules. Buildings are analyzed and layouts are generated through communication and interaction between these two systems. The benefit of using an interactive system is that the complementary properties of the two schemes are employed to strengthen the overall process. Ontologies capture knowledge relating to architectural design principles, building anatomy, structure, and systems. Shape grammar rules embody change through geometric manipulation and transformation. Existing buildings are analyzed using this approach, and three-dimensional models are automatically generated. Two particular building types, the vernacular rowhouse and high-rise apartment building, both from Baltimore, Maryland, are presented to illustrate the process and for comparing the utilized methodologies.

The paper introduces and discusses the notion of decomposition of a configuration problem within the framework of a structured logical approach. The paper describes under which conditions a given configuration problem can be decomposed into a set of noninteracting subproblems and how to exploit such a decomposition, both for improving the performance of the configurator and for supporting interactive configuration. Different kinds of decomposition are considered, but all of them exploit, as much as possible, the explicit representation of the partonomic relations in the language, a KL-One like representation formalism augmented with constraints for expressing complex interrole relations. The paper introduces a notion of boundness among constraints, which is used for formally specifying different types of decomposition. One decomposition strategy aims at singling out the components and subcomponents that are directly related to the constraints put by the user's requirements; the configurator exploits such decomposition by first configuring that portion of the product and then configuring the parts that are not related to the user's requirements. Another decomposition strategy verifies whether the set of constraints for the product to be configured can be split into a set of noninteracting problems. In such a case the configurator solves the configuration problem by splitting the whole search space into a set of smaller search spaces. Different combinations of these two decomposition techniques are considered, and the impact of the decomposition strategies on the performance of the configurator is evaluated via a set of experiments using the configuration of computer systems as a test bed. The results of the experiments show a significant reduction of the computational effort (both in terms of number of backtrackings and in CPU time) when decomposition strategies are used.

Biomimetic design applies biological analogies to solve design problems and has been known to produce innovative solutions. However, when designers are asked to perform biomimetic design, they often have difficulty recognizing analogies between design problems and biological phenomena. Therefore, this research aims to investigate designer behaviors that either hinder or promote the use of analogies in biomimetic design. A verbal protocol study was conducted on 30 engineering students working in small teams while participating in biomimetic design sessions. A coding scheme was developed to analyze cognitive processes involved in biomimetic design. We observed that teams were less likely to apply overall biological analogies if they tended to recall existing solutions that could be easily associated with specific superficial or functional characteristics of biological phenomena. We also found that the tendency to evaluate ideas, which reflects critical thinking, correlates with the likelihood of identifying overall biological analogies. Insights from this paper may contribute toward developing generalized methods to facilitate biomimetic design.

The tool switching problem (ToSP) is well known in the domain of flexible manufacturing systems. Given a reconfigurable machine, the ToSP amounts to scheduling a collection of jobs on this machine (each of them requiring a different set of tools to be completed), as well as the tools to be loaded/unloaded at each step to process these jobs, such that the total number of tool switches is minimized. Different exact and heuristic methods have been defined to deal with this problem. In this work, we focus on memetic approaches to this problem. To this end, we have considered a number of variants of three different local search techniques (hill climbing, tabu search, and simulated annealing), and embedded them in a permutational evolutionary algorithm. It is shown that the memetic algorithm endowed with steepest ascent hill climbing search yields the best results, performing synergistically better than its stand-alone constituents, and providing better results than the rest of the algorithms (including those returned by an effective ad hoc beam search heuristic defined in the literature for this problem).

This paper examines the architecture of back-propagation neural networks used as approximators by addressing the interrelationship between the number of training pairs and the number of input, output, and hidden layer nodes required for a good approximation. It concentrates on nets with an input layer, one hidden layer, and one output layer. It shows that many of the currently proposed schemes for selecting network architecture for such nets are deficient. It demonstrates in numerous examples that overdetermined neural networks tend to give good approximations over a region of interest, while underdetermined networks give approximations which can satisfy the training pairs but may give poor approximations over that region of interest. A scheme is presented that adjusts the number of hidden layer nodes in a neural network so as to give an overdetermined approximation. The advantages and disadvantages of using multiple output nodes are discussed. Guidelines for selecting the number of output nodes are presented.

This paper explores using Shannon's entropy of information to measure linkographs of 12 design sessions that involved six architects in two different experimental conditions. The aim is to find a quantitative tool to interpret the linkographs. This study examines if the differences in the design processes and the design outcomes can be reflected in the entropic interpretations. The results show that the overall entropy of one design condition is slightly higher than the other. Further, there are indications that the change of entropy might reflect design outcomes.

Data-enabled design (DED) is a promising new methodology for designing with users from within their own context in an iterative and hands-on fashion. However, the agile and flexible qualities of the methodology do not directly translate to every context. In this article, we reflect on the design process of an intelligent ecosystem, called ORBIT, and a proposed evaluative study planned with it. This was part of a DED project in collaboration with a medical hospital to study the post-operative behavior in the (remote) context of bariatric patients. The design and preparation of this project and the process towards an eventual study rejection from the medical ethical committee (METC) provide rich insights into (1) what it means to conduct DED research in a clinical context, and (2) where the boundaries of the method might lie in this specific application area. We highlight insights from carefully designing the substantial infrastructure for the study, and how different aspects of DED translated less easily to the clinical context. We analyze the proposed study setup through the lenses of several modifications we made to DED and further reflect on how to expand and scale up the methodology and adapt the process for the clinical context.

Understanding user perceptions of interacting with the virtual world is one of the research focuses in recent years, given the rapid proliferation of virtual reality (VR) and driven to establish the metaverse. Users can generate a familiar connection between their bodies and the virtual world by being embodied in virtual hands, and hand representations can induce users’ embodiment in VR. The sense of embodiment represents the cognitive awareness of one's manifestation and includes three subcomponents: the sense of body ownership, agency and self-location. There is insufficient evidence in the literature about the effects of hand designs on the embodiment, especially based on studying its three subcomponents. This study investigates how virtual hand designs with five realism levels influence the three subcomponents of embodiment in VR. This research employs a self-report questionnaire commonly used in the literature to assess embodiment and evaluates agency and self-location by introducing implicit methods (intentional binding and proprioceptive measurement) derived from psychology. Besides, the objective data of eye tracking is used to explore the connection between embodiment and hand designs, and classifying participants’ eye tracking data to help analyze the link between embodiment and user attention. Overall, this research makes a major contribution through a systematic exploration of users’ embodied experience in VR and offers important evidence of the effects of virtual hand designs on body ownership, agency, and self-location, respectively. In addition, this study provides a valuable reference for further investigation of embodiment through implicit and objective methods, and practical design recommendations for virtual hand design in VR applications.

The need to model and to reason about design alternatives throughout the design process demands robust representation schemes of function, behavior, and structure. Function describes the physical effect imposed on an energy or material flow by a design entity without regard for the working principles or physical solutions used to accomplish this effect. Behaviors are the physical events associated with a physical artifact (or hypothesized concept) over time (or simulated time) as perceived by an observer. Structure, the most tangible concept, partitions an artifact into meaningful constituents such as features, Wirk elements, and interfaces in addition to the widely used assemblies and components. The focus of this work is on defining a model for function-based representations that can be used across various design methodologies and for a variety of design tasks throughout all stages of the design process. In particular, the mapping between function and structure is explored and, to a lesser extent, its impact on behavior is noted. Clearly, the issues of a function-based representation's composition and mappings directly impact certain computational synthesis methods that rely on (digitally) archived product design knowledge. Moreover, functions have already been related to not only form, but also information of user actions, performance parameters in the form of equations, and failure mode data. It is essential to understand the composition and mappings of functions and their relation to design activities because this information is part of the foundation for function-based methods, and consequently dictates the performance of those methods. Toward this end, the important findings of this work include a formalism for two aspects of function-based representations (composition and mappings), the supported design activities of the model for function-based representations, and examples of how computational design methods benefit from this formalism.

Probabilistic and stochastic algorithms have been used to solve many hard optimization problems since they can provide solutions to problems where often standard algorithms have failed. These algorithms basically search through a space of potential solutions using randomness as a major factor to make decisions. In this research, the knapsack problem (optimization problem) is solved using a genetic algorithm approach. Subsequently, comparisons are made with a greedy method and a heuristic algorithm. The knapsack problem is recognized to be NP-hard. Genetic algorithms are among search procedures based on natural selection and natural genetics. They randomly create an initial population of individuals. Then, they use genetic operators to yield new offspring. In this research, a genetic algorithm is used to solve the 0/1 knapsack problem. Special consideration is given to the penalty function where constant and self-adaptive penalty functions are adopted.

This paper presents the requirements and needs for establishing a benchmarking protocol that considers representation characteristics, supported cognitive criteria, and enabled reasoning activities for the systematic comparison of function modeling representations. Problem types are defined as reverse engineering, familiar products, novel products, and single-component systems. As different modeling approaches share elements, a comparison of modeling approaches on multiple levels was also undertaken. It is recommended that researchers and developers of function modeling representations collaborate to define a canonically acceptable set of benchmark tests and evaluations so that clear benefits and weaknesses for the disparate collection of approaches can be compared. This paper is written as a call to action for the research community to begin establishing a benchmarking standard protocol for function modeling comparison purposes. This protocol should be refined with input from developers of the competing approaches in an academically open environment. At the same time, the benchmarking criteria identified should also serve as a guide for validating a modeling approach or analyzing its failure.

This paper examines the need for explicit representations of the design space, in response to Woodbury and Burrow. Specifically, their proposal for a particular search strategy, by means of which one can reuse past experiences explicitly represented by previously traversed paths, is examined. This is done by exploring issues with respect to design search and representation in general, while relating these to specific issues raised by Woodbury and Burrow. The paper concludes by suggesting that their arguments essentially point to devising an appropriate “programming language” for design.

In their quest to manage the complexity of offering greater product variety, firms in many industries are considering platform-based development of product families. Key in this approach is the sharing of components, modules, and other assets across a family of products. Current research indicates that companies are often choosing physical elements of the product architecture (i.e., components, modules, building blocks) for building platform-based product families. Other sources for platform potential are widely neglected. We argue that for complex products and systems with hierarchic product architectures and considerable freedom in design, a new platform type, the system layout, offers important commonality potential. This layout platform standardizes the arrangement of subsystems within the product family. This paper is based on three industry case studies, where a product family based on a common layout could be defined. In combination with segment-specific variety restrictions, this results in an effective, efficient, and flexible positioning of a company's products. The employment of layout platforms leads to substantial complexity reduction, and is the basis for competitive advantage, as it imposes a dominant design on a product family, improves its configurability, and supports effective market segmentation.

Defining the aesthetic and emotional value of a product is an important consideration for its design. Furthermore, if several designers are faced with the task of creating an object that describes a certain emotion/perception (aggressive, soft, heavy, etc.), each is most likely to interpret the emotion/perception with different shapes composed of a set of different geometric features. The authors propose an automatic approach to formalize the relationships between geometric information of three-dimensional objects and the intended emotional content using fuzzy logic. In addition, the automatically generated fuzzy knowledge base was compared to the user's perceptions and to the manually constructed fuzzy knowledge base. The initial findings indicate that the approach is valid to formalize geometric information with perceptions and validate the author's manually developed fuzzy models.

Analyses of styles in design have paid little attention to how people see style and how designers use perceptions of style to guide designing. Although formal and computational methods for analyzing styles and generating designs provide impressively parsimonious accounts of what some styles are, they do not address many of the factors that influence how humans understand styles. The subtlety of human style judgments raises challenges for computational approaches to style. This paper differentiates between a range of distinct meanings of “style” and explores how designers and ordinary people learn and apply perceptual similarity classes and style concepts in different situations to interpret and create designed artifacts. A range of psychological evidence indicates that style perception is dependent on knowledge and involves the interaction of perceptual recognition of style features and explanatory inference processes that create a coherent understanding of an object as an exemplar of a style. This article concludes by outlining how formal style analyses can be used in combination with psychological research to develop a fuller understanding of style perception and creative design.