This paper addresses manufacturing research involving contributions in computer science, control theory and material processing science. The research objective has been to develop intelligent, †self-directed‡ control systems which enable in situ control path generation based on †product-processing‡ feedback. A †product-directed‡ control philosophy which emphasizes product quality is described together with a generic architecture for representing process knowledge. Specific details are presented involving the development and application of a self-directed material processing system for autoclave curing of composites together with recent research results, production challenges and future research in †self-improving‡ material processing systems.

A general constraint-based formulation for the analysis and partial synthesis of two-dimensional truss structures is presented. This formulation is general in that it handles statically determinate and statically indeterminate trusses with pin and roller supports, and concentrated joint loads. The formulation is constraint-based in that the physical behavior of truss components is declaratively represented using constraints.

The analysis and partial synthesis of a truss structure manifest themselves in proving the satisfiability of the constraints associated with the structural components. An artificial intelligence approach called constraint logic programming is used for representing and satisfying constraints. A constraint logic programming language, called CLP(R), is used for implementing the formulation. The implemented program consists of sixteen rules. These rules are used for both the analysis and partial synthesis of truss structures. Several truss analysis and synthesis examples using the formulation are presented.

Genetic algorithm (GA) and singular value decomposition (SVD) are deployed for the optimal design of both Gaussian membership functions of antecedents and the vector of linear coefficients of consequents, respectively, of adaptive neurofuzzy inference systems (ANFIS) networks that are used for modeling of the explosive cutting process of plates by shaped charges. The aim of such modeling is to show how the depth of penetration varies with the variation of important parameters, namely, the apex angle, standoff, liner thickness, and mass of charge. It is demonstrated that SVD can be effectively used to optimally find the vector of linear coefficients of conclusion parts in ANFIS models and their Gaussian membership functions in premise parts are determined by a GA.

This paper aims to situate functional abstraction in light of systems thinking. While function does not extensively appear in systems thinking literature, the literature does identify function as part of systems thinking that enables us to recognize and connect that function has a role in building a systems thinking approach for students. A systems thinking approach is valuable for students since it helps them view a system holistically. In this research, we measure how well students are able to abstract function. We asked students to generate functions for two different products and examined how students taught functional modeling and function enumeration compare to students who are only taught function enumeration. The student responses were examined using a rubric that we developed and validated for assessing function. This rubric may be used to classify functions by correctness (correct, partially correct, and incorrect) and categories (high level, interface, low level, and ambiguous). On questions where students were not explicitly asked to write a high-level function or low-level function, and so on, students who were taught functional modeling were able to better demonstrate systems thinking in their responses (low-level and interface functions) than those students who were only taught function enumeration.

Methods for efficient variance-based global sensitivity analysis of complex high-dimensional problems are presented and compared. Variance decomposition methods rank inputs according to Sobol indices that can be computationally expensive to evaluate. Main and interaction effect Sobol indices can be computed analytically in the Kennedy and O'Hagan framework with Gaussian processes. These methods use the high-dimensional model representation concept for variance decomposition that presents a unique model representation when inputs are uncorrelated. However, when the inputs are correlated, multiple model representations may be possible leading to ambiguous sensitivity ranking with Sobol indices. In this work, we present the effect of input correlation on sensitivity analysis and discuss the methods presented by Li and Rabitz in the context of Kennedy and O'Hagan's framework with Gaussian processes. Results are demonstrated on simulated and real problems for correlated and uncorrelated inputs and demonstrate the utility of variance decomposition methods for sensitivity analysis.

Configuration is a complex task generally involving varying measures of constraint satisfaction, optimization, and the management of soft constraints. Although many successful systems have been developed, these are often difficult to maintain and to generalize in rapidly changing domains. In this paper, we consider building intelligent knowledge-based systems with maintainability well to the fore in our requirements for such systems. We introduce two case studies: the initial proof of concept, which was in the domain of computer configuration, and a further field-tested study, the configuration of compressors. Central to our approach is the use of the proof planning technique, and the clean separation of different kinds of knowledge: factual, heuristic, and strategic.

Artificial Intelligence has been very active in developing high-level symbolic reasoning paradigms that have resulted in practical expert systems. However, with a few exceptions, it has paid little attention to the automation of spatial reasoning. On the other hand, spatial reasoning has attracted the interest of several researchers in Robotics. One of the important problems that have been investigated is motion planning, and very significant results have been obtained. This paper describes an implemented system for designing pipe layouts automatically using motion planning techniques. It introduces a new approach to pipe layout design automation in which pipe routes are treated as trajectories left behind by rigid objects (‘robots’). We have implemented this approach in a basic Pipe Router that is described in detail in this paper. We have extended this router in order to make it capable of treating a variety of other constraints which are typical of practical pipe layout design problems. These constraints relate to the process carried out in the pipes, to the design of their mechanical support, and to the constructability and the ease of operation and maintenance of the designed pipe systems.

This paper claims that style, in addition to being identified by common visible physical characteristics of form, can be thought of in terms of a set of common abstract characteristics. A prototype computational design support tool is described that explores this idea in the domain of architecture. The Architect's Collaborator (TAC) supports articulation and evaluation of abstract characteristics of style (e.g., experiential characteristics such as privacy and shelter) and does so by mapping abstract characteristics to details of physical form. The implementation of TAC is described and successful experiments are reported in which abstract characteristics of Frank Lloyd Wright's Prairie houses were mapped to physical form characteristics and used to evaluate Prairie and non-Prairie houses.

We present a methodology of learning fuzzy rules using an iterative genetic algorithm (GA). The approach incorporates a scheme of partitioning the entire solution space into individual subspaces. It then employs a mechanism to progressively relax or tighten the constraint. The relaxation or tightening of constraint guides the GA to the subspace for further iteration. The system referred to as the iterative GA learning module is useful for learning an efficient fuzzy control algorithm based on a predefined linguistic terms set. The overall approach was applied to learn a fuzzy algorithm for a water bath temperature control. The simulation results demonstrate the effectiveness of the approach in automating an industrial process.

This paper is concerned with presenting guidelines to aide in the selection of the appropriate network architecture for back-propagation neural networks used as approximators. In particular, its goal is to indicate under what circumstances neural networks should have two hidden layers and under what circumstances they should have one hidden layer. Networks with one and with two hidden layers were used to approximate numerous test functions. Guidelines were developed from the results of these investigations.

This position paper proposes the P5BL initiative and vision, that is, Problem-, Project-, Product-, Process-, People-Based Learning. A definition is proposed for the P5BL teaching and learning approach. The discussion identifies key pedagogical issues and innovative roles of computing for each of the five critical Ps in a multisite, cross-disciplinary, project-centered, and team-oriented Architecture/Engineering/Construction environment.

This paper interprets the concept of biologically inspired design as understanding design based on the biological evidence. Borrowing its concept of design competence from Chomsky's definition of linguistic competence, the paper reviews biological evidence from fields including evolution, genetics, and animal behavior from the perspective of design research to propose that design competence is the product of an evolutionary history during which five key developments in cognitive evolution came together: conception unbounded by sensory perception, symbolic manipulation at a level of metarepresentation, theory of mind, curiosity, and mental time travel. These cognitive capabilities were derived from the biological evidence based upon the criteria that they are presumed to be unique to humans (Homo sapiens), they may be lost because of neurodegenerative diseases or they may fail to develop because of neurodevelopmental disorders, and they are not immediately present upon birth and develop as a child's brain matures. Based on these five capabilities, the paper concludes by discussing how computation may provide a useful way to understand the origins and evolution of design competence.

In engineering design, the end goal is the creation of an artifact, product, system, or process that fulfills some functional requirements at some desired level of performance. As such, knowledge of functionality is essential in a wide variety of tasks in engineering activities, including modeling, generation, modification, visualization, explanation, evaluation, diagnosis, and repair of these artifacts and processes. A formal representation of functionality is essential for supporting any of these activities on computers. The goal of Parts 1 and 2 of this Special Issue is to bring together the state of knowledge of representing functionality in engineering applications from both the engineering and the artificial intelligence (AI) research communities.

Manufacturing knowledge is maintained primarily in the unstructured text in industry. To facilitate the reuse of the knowledge, previous efforts have utilized Natural Language Processing (NLP) to classify manufacturing documents or to extract structured knowledge (e.g. ontology) from manufacturing text. On the other hand, extracting more complex knowledge, such as manufacturing rule, has not been feasible in a practical scenario, as standard NLP techniques cannot address the input text that needs validation. Specifically, if the input text contains the information irrelevant to the rule-definition or semantically invalid expression, standard NLP techniques cannot selectively derive precise information for the extraction of the desired formal manufacturing rule. To address the gap, we developed the feedback generation method based on Constraint-based Modeling (CBM) coupled with NLP and domain ontology, designed to support formal manufacturing rule extraction. Specifically, the developed method identifies the necessity of input text validation based on the predefined constraints and provides the relevant feedback to help the user modify the input text, so that the desired rule can be extracted. We proved the feasibility of the method by extending the previously implemented formal rule extraction framework. The effectiveness of the method is demonstrated by enabling the extraction of correct manufacturing rules from all the cases that need input text validation, about 30% of the dataset, after modifying the input text based on the feedback. We expect the feedback generation method will contribute to the adoption of semantics-based technology in the manufacturing field, by facilitating precise knowledge acquisition from manufacturing-related documents in a practical scenario.

Detecting and recognizing objects is one of the most important uses of vision systems in nature and is consequently highly evolved. This paper aims to accurately detect an object using its shape and color information from a complex background. In particular, we evaluated our algorithm to detect 19 different integrated circuits (IC) from 10 different printed circuit boards (PCB) of different colors. We have compared three different shape descriptors for four different color space models. We have evaluated shape detection algorithms in different lighting conditions (indoor, outdoor, and controlled light source) to find suitable illumination for image acquisition. We undertook statistical hypothesis testing to find the effect of color space models and shape descriptors on the accuracy, false positive and false negative rates. While measuring accuracy, we have noted that L*a*b* color space is significantly worse, and the best result is obtained in YCbCr color space using bounding box shape descriptors for 2500 Lux using LED.

One of the fundamental elements that determines the precision of coordinate measuring machines (CMMs) is the probe, which locates measuring points within measurement volume. In this paper genetically generated fuzzy knowledge based models of three-dimensional (3-D) probing accuracy for one- and two-stage touch trigger probes are proposed. The fuzzy models are automatically generated using a dedicated genetic algorithm developed by the authors. The algorithm uses hybrid coding, binary for the rule base and real for the database. This hybrid coding, used with a set of specialized operators of reproduction, proved to be an effective learning environment in this case. Data collection of the measured objects' coordinates was carried out using a special setup for probe testing. The authors used a novel method that applies a low-force high-resolution displacement transducer for probe error examination in 3-D space outside the CMM measurement. The genetically generated fuzzy models are constructed for both one stage (TP6) and two stage (TP200) types of probes. First, the optimal number of settings is defined using an analysis of the influence of fuzzy rules on TP6 accuracy. Then, once the number of settings is obtained, near optimal fuzzy knowledge bases are generated for both TP6 and TP200 triggering probes, followed by analysis of the finalized fuzzy rules bases for knowledge extraction about the relationships between physical setup values and error levels of the probes. The number of fuzzy sets on each premise leads to the number of physical setups needed to get satisfactory error profiles, whereas the fuzzy rules base adds to the knowledge linking the design experiment parameters to the pretravel error of CMM machines. Satisfactory fuzzy logic equivalents of the 3-D error profiles were obtained for both TP6 and TP200 with root mean squsre errors ranging from 0.00 mm to a maximum of 0.58 mm.

This paper proposes and evaluates a neural network-based method for simulating manufacturing processes that exhibit both noncontinuous and stochastic behavior processes more conventionally modeled, using discrete-event simulation algorithms. The incentive for developing the technique is its potential for rapid execution of a simulation through parallel processing, and facilitation of the development and improvement of models particularly where there is limited theory describing the dependence between component processes. A brief introduction is provided to a radial-Gaussian neural network architecture and training process, the system adopted for the work presented in this paper. A description of the basic approach proposed for applying this technology to simulation is then described. This involves the use of a modularized neural network approach to model construction and the prediction of the occurrence of events using information retained from several previous states of the simulation. A class of earth-moving systems, comprising a push-dozer and a fleet of scrapers, is used as the basis for assessing the viability and performance of the proposed approach. A series of experiments show the neural network to be capable of both capturing the characteristic behavior and making an accurate prediction of production rates of scraper-based earth-moving systems. The paper concludes with an indication of some areas for further development and evaluation of the technique.

Why has integration and interoperability of design and manufacturing knowledge proven so difficult? There is a common puzzle that appears in your Sunday newspaper or The New Yorker magazine.

Linkage-type mechanisms have numerous applications in industry especially for automation Unfortunately, they are less popular due to lack of proper design tools. This paper describes our efforts to remove the technological barrier in mechanisms design automation. Although the ideas presented apply to automation of mechanisms design in general, the paper discusses the development of an expert system for a particular sub-set of mechanisms called Dwell mechanisms.

Many essential and desirable motion characteristics of mechanisms are so implicit that they are difficult to control by analytical methods. By systematically and extensively studying the entire motion characteristics of hundreds of linkages, a comprehensive classification system and heuristics were developed. This qualitative classification scheme led to a finite set of linkage models that cover the entire design space in the sense that any possible design falls under one or more of the models. Our system, called Dwell-Expert, incorporates this design expertise to select the best linkage model for a given set of design specifications and to compare that model against alternatives. The new design methodology and its implementation in AGNESS (A Generalized Network-based Expert System Shell) are explained. A design example is also presented. Our system can reduce even an experienced designer’s initial-design time from a day or more to a minute or less, assuming specifications have already been formulated. Such results motivate extension of this design methodology to other areas of mechanical design and engineering design in general.

This article is an attempt to improve the efficiency of procedures for compositional synthesis of design solutions using building blocks. These procedures have found use in a wide range of applications, and are one of the most substantial outcomes of research into automated synthesis of design solutions. Due to their combinatorial nature, these procedures are highly inefficient in solving problems, especially when the database of building blocks for synthesis or the problem size is large. Previous literature often focuses on improving only the algorithm part of a procedure, although it is both its algorithm and database which together determine the overall efficiency of the procedure. This article reports the construction and analysis of an improved algorithm, based on bidirectional search, for efficient compositional synthesis of design solutions using a set of building blocks.