Direct planning of assembly sequences is either very complicated computationally, or requires an ‘expert’ user. This paper presents a new approach to the automatic generation of assembly plans which involves two phases. An ‘exploded’ layout of the assembled product is first found from its topological and geometric description by a graph-based procedure. Then, a knowledge-based technique is used for planning of selected assembly sequences by utilizing theoretical as well as heuristic knowledge of mechanical components and assembly processes. Besides serving the assembly planning phase, automatic generation of exploded-views could also be a most desirable feature of any computer-aided design system. An example demonstrates the application of the method to an actual product.

Protective packaging buffer design is a complex process which involves manipulating both analytical and graphical data, and expert knowledge. The design activities of this process can be represented as a series of procedural tasks, and it is therefore suitable for design automation efforts. Currently, the design activities of the protective packaging buffers are predominantly manual and done on a trial and error basis. In this paper, PROPACK, a knowledge-based CAD system which has been developed to automate this design process will be discussed. A description of how the knowledge and information in PROPACK are structured and stored in the database will also be made. Preliminary tests have found that PROPACK is able to generate possible design solutions that either require external or simple internal cushioning features used for protection. These results have found to compare quite favourably with the BASF worked examples.

This research originates in the work started several years ago at Worcester Polytechnic Institute dedicated to the investigation, modelling and evaluation of multiagent based design. The main thrust behind our approach is the idea of finding elementary patterns of agent problem-solving and interaction in design tasks. To achieve this goal we introduced and defined the concept of Single Function Agents, (SiFAs) (Dunskus, 1995; SiFA, 1995). SiFAs are agents specialized to perform one single generic function during the design process. Some typical functions would be selection, evaluation, and critique. These types of agents can be instantiated for different, particular design domains.

Artificial intelligence (AI) applications to design have tended to focus on modeling and automating aspects of single discipline design tasks. Relatively little attention has thus far been devoted to representing the kinds of design ‘metaknowledge’ needed to manage the important interface issues that arise in concurrent design, that is, multidisciplinary design decision-making. This paper provides a view of the process and management of concurrent design and evaluates the potential of two AI approaches—blackboard architectures and co-operative distributed problem-solving (CDPS)—to model and support the concurrent design of complex artifacts. A discussion of the process of multidisciplinary design highlights elements of both sequential and concurrent design decision-making. We identify several kinds of design metaknowledge used by expert managers to: partition the design task for efficient execution by specialists; set appropriate levels of design conservatism for key subsystem specifications; evaluate, limit and selectively communicate design changes across discipline boundaries; and control the sequence and timing of the key (highly constrained and constraining) design decisions for a given type of artifact. We explore the extent to which blackboard and CDPS architectures can provide valid models of and potential decision support for concurrent design by (1) representing design management metaknowledge, and (2) using it to enhance both horizontal (interdisciplinary) and vertical (project life cycle) integration among product design, manufacturing and operations specialists.

Sometime in the mid-1980s, in an urban restaurant some-where in the United States, I was sitting next to a Japanese engineer who was talking to a U.S. engineer. The following is the record of their conversation that I made with my Sony Walkman.

In this paper, a methodology for inducing trends in a first principle reasoning system for design innovation is presented. Dimensional Variable Expansion is used in 1stPRINCE (FIRST PRINciple Computational Evaluator) to create additional design variables and introduce new prototypes. Trends are observed at each generation of the prototype and induction is used to predict optimal constraint activity at the limit of the iterative procedure. The inductive mechanism is applied to a constant-radius beam under flexural load and a tapered beam of varying radius and superior performance is derived. A circular wheel is created from a primitive-prototype consisting of a rectangular, spinning block that is optimized for minimum resistance to spinning. Although presented as a technique to perform innovative design, the inductive methodology can also be utilized as an AI approach to shape optimization.

This paper presents an approach for the case adaptation, especially case repairing, in a case-based process planning system: PROCASE, for machining of rotational parts. In PROCASE, a new process plan is generated by adapting an existing similar process plan from its case library. Case adaptation is a crucial issue in achieving an automated case-based process planning system. This is because, usually, an existing process plan cannot necessarily produce an exact identical part as of the desired part. Adaptation is essential to tailor this existing plan to generate a new process plan for the new part. The case adaptation in this paper comprises case modification, case simulation, and case repairing. The modifier uses the knowledge extracted from case library to edit the retrieved similar plan. The simulator plays an important role in verifying the adapted plan as well as in directing the plan repairing. The repairing rules are indexed by the error messages obtained from the simulation. With the proposed case adaptation, the system will have the capability to repair the erroneous plans to achieve an automated and intelligent process planning system. This paper will first briefly introduce the case representation and case retrieval in PROCASE. Then the rest of the paper is dedicated to the case adaptation.

Physical systems can be modelled at many levels of approximation. The right model depends on the problem to be solved. In many cases, a combination of models will be more effective than a single model. Our research investigates this idea in the context of engineering design optimization. We present a family of strategies that use multiple models for unconstrained optimization of engineering designs. The strategies are useful when multiple approximations of an objective function can be implemented by compositional modelling techniques. We show how a compositional modelling library can be used to construct a variety of locally calibratable approximation schemes that can be incorporated into the optimization strategies. We analyze the optimization strategies and approximation schemes to formulate and prove sufficient conditions for correctness and convergence. We also report experimental tests of our methods in the domain of sailing yacht design. Our results demonstrate dramatic reductions in the CPU time required for optimization, on the problems we tested, with no significant loss in design quality.

This paper discusses a work in progress in the development of computer tools for qualitative modeling analysis and evaluation of conceptual structural designs. In the conceptual design stage the description of a structure is incomplete and imprecise, and does not permit the use of traditional numerical analysis tools. We describe a prototype system, QLRS, for qualitative evaluation of lateral load resistance in frames. The primary goal of the evaluation of structural response is to identify undesirable structural behavior. In QLRS, the evaluation process consists of three basic tasks. (1) identification of the story and structure models comprising the lateral load resisting system. We term this task structural system interpretation. (2) Qualitative analysis of the story and structure models, and approximate evaluation of the story drifts. We term this task structural behavior interpretation. (3) Assessment of the performance of the lateral load resisting system, in which the results of the structural system interpretation and the structural behavior interpretation are compared against the requirements for complete load path and relative story drift. Currently, QLRS is able to reason about load path discontinuities and soft-story behavior problems in 2-D moment resisting frames.

Design knowledge is continually refined and expanded through experience. This research is concerned with design knowledge expressed as constraints. A simple learning mechanism simulates an expert designer's ability to incrementally adjust her knowledge when presented with slightly new problems. In response to unsatisfied expectations during the design process the system will examine its general knowledge about the design artifact, discover some relevant constraining knowledge, and convert that knowledge into a design constraint for future use. This process, referred to as constraint inheritance, should automatically improve the problem-solving performance.

Process planning is the function that converts an engineering design into a manufacturing plan. One of the problems in feature-based process planning is the sequencing of features. Features must be given an order for removal. This order, or sequence, is partially dependent on the geometric relationships between the features. If the geometric relationships between features are such that they dictate a particular sequence, the features are said to have an interaction. Identifying these interactions is an important first step in creating the process plan. An approach to solve this problem using constructive solid geometry operations and the Episodal Associative Memory (EAM) is demonstrated. The EAM is an associative memory that integrates dynamic memory organization and neural computing techniques. The geometric feature relationships can be represented by a pattern. This pattern captures very qualitative information about the geometric positions fo the features. The EAM can organize these patterns into groups of similar geometric relationships. A method for dealing with exceptions, and for retrieving and storing general machining problems associated with interacting features will be described. The system implemented is shown to correctly sequence several types of feature interactions.

This article presents an assessment of four management systems to expose the essential characteristics of each management system, including planning techniques, problem representation, concurrency, and flexibility. The experimental part of the research shows that existing management systems can be used to attack a variety of problems. The authors conclude that flexible planning systems are quite beneficial since they can be used to solve a variety of design problems by making small modifications in the definition of their tools.

The spatial synthesis problem addressed in this paper is the configuration of rectangles in 2D space, where the sides of the rectangles are parallel to an orthogonal coordinate system. Variables are the locations of the edges of the rectangles and their orientations. Algebraic constraints on these variables define a layout and constitute a constraint satisfaction problem. We give a new O(n2) algorithm for incremental path-consistency, which is applied after adding each algebraic constraint. Problem requirements are formulated as spatial relations between the rectangles, for example, adjacency, minimum distance, and nonoverlap. Spatial relations are expressed by Boolean combinations of the algebraic constraints; called disjunctive constraints. Solutions are generated by backtracking search, which selects a disjunctive constraint and instantiates its disjuncts. The selected disjuncts describe an equivalence class of configurations that is a significantly different solution. This method generates the set of significantly different solutions that satisfy all the requirements. The order of instantiating disjunctive constraints is critical for search efficiency. It is determined dynamically at each search state, using functions of heuristic measures called textures. Textures implement fail-first and prune-early strategies. Extensions to the model, that is, 3D configurations, configurations of nonrectangular shapes, constraint relaxation, optimization, and adding new rectangles during search are discussed.