Abstract: A method is proposed for decomposing large optimization problems encountered in the design of engineering systems such as an aircraft into a number of smaller subproblems. The decomposition is achieved by organizing the problem and the subordinated subproblems in a tree hierarchy and optimizing each subsystem separately. Coupling of the subproblems is accounted for by subsequent optimization of the entire system based on sensitivities of the suboptimization problem solutions at each level of the tree to variables of the next higher level. A formalization of the procedure suitable for computer implementation is developed and the state of readiness of the implementation building blocks is reviewed showing that the ingredients for the development are on the shelf. The decomposition method is also shown to be compatible with the natural human organization of the design process of engineering systems. The method is also examined with respect to the trends in computer hardware and software progress to point out that its efficiency can be amplified by network computing using parallel processors.

Abstract: ' Introduction T concept of optimization is intrinsically tied to natural phenomena as well as to the human desire to excel. Sir George Cayley (1773-1857) measured the shape of a trout and noted, without mathematical proof, that the trout was ideally proportioned to minimize flow resistance. Theodore von Kdrmdn observed that this is precisely the shape of a lowdrag airfoil. Oliver Wendell Holmes (1809-1894), in his classic verse, "The Deacon's Masterpiece; or, The Wonderful OneHoss Shay," recorded man's desire to produce a uniformly strong, durable product. In this case it was the structural design of a shay to last a hundred years. Perhaps the first analytical work in structural optimization was by Maxwell in 1869, followed by the better-known work of Michell in 1904. These works provided theoretical lower bounds on the weight of trusses, and, although highly idealized, offer considerable insight into the structural optimization problem and the design process. The 1940s and early 1950s saw development of component optimization in such works as Shanley's Weight-Strength Analysis of Aircraft Structures. Also during this period, availability of the digital computer led to application of linear programming techniques to plastic design of frames, for example, the work of Heyman. This early numerical work is particularly significant in that it used mathematical programming techniques developed in the operations research community to solve structural design problems. Schmit in 1960 was the first to offer a comprehensive statement of the use of mathematical programming techniques to solve the nonlinear-inequality-constrained problem of designing elastic structures under a multiplicity of loading conditions. This work is significant, not only in that it ushered in an era of structural optimization, but also because it offered a new philosophy of engineering design which is only now beginning to be broadly applied. In Ref. 9 Schmit provides an excellent historical review of the development of this concept. Although this discussion will emphasize numerical design techniques, it is important to note that there has been an extensive amount of research in analytical methods of design. That work, although sometimes lacking the practicality of being applied to realistic structures, is nonetheless of fundamental importance because it provides insight into the design problem and because it often provides theoretical lower bounds against which more practical designs may be judged. References 10 and 11 provide an extensive review of the state-of-the-art in analytical design techniques. It is the use of numerical techniques in structural optimization that is emphasized here. The purpose is not to offer a tutorial on optimization or a comprehensive literature survey, although such works are referenced. Rather, it is to look briefly at the short history of modern structural optimization and assess the state-of-the-art from a somewhat more philosophical viewpoint. In this way we may begin to understand the ramifications of this fascinating approach to design. By learning what is now possible and what is not now possible, we may encourage the use of these techniques by practicing designers as well as identify research and development needs of the future.

Abstract: This paper outlines the strategies adopted by the psychologists and ergonomists of the HUSAT Research Group to help organizations learn about information technology and systematically review and plan its organizational ramifications. An essential point is that it is not enough to understand the technology; effective implementation demands the ability to establish organizational needs and to choose a form of technology which will meet them. The paper examines three ways of designing systems. Firstly, a technology-led approach which leads to 'fire fighting' when the negative organizational effects become apparent. A second method has tried to compensate for this by involving users in the design process. Unfortunately by the time the users have come to terms with their new task and are able to make a contribution, the system has usually been designed. The third method of design expressly seeks to give users the time and opportunity to learn how to contribute to design, by making the design process e...

Abstract: Significantly new challenges were presented for the design of the 3081 Processor Unit since it was the first IBM large system implemented in LSI technology. Solutions had to be found for a new set of problems in order to achieve the required product objectives while maintaining an acceptable development cost and schedule. In this paper, the design aspects and the characteristics of the 3081 Processor Unit are described and the tradeoffs that were made due to the implementation of LSI are presented. A design strategy was chosen that included tradeoffs covering the areas of machine organization, performance level, implementation costs, testing and servicing aids, and development schedules. An innovative verification effort was introduced into the design process, capitalizing on a hardware flowcharting discipline and rigorous design rules. On the basis of this development experience, some thoughts concerning VLSI implementations are explored.

Abstract: The emphasis of most computer aided design is on the benefits to the user of the CAD system rather than on the possible benefits to the user of the resulting end products. This paper looks at some of the problems for any designer using a CAD system and in particular at the problems of involving the end user of a product directly in the design process. A simplified three-dimensional computer input device is described as an aid to the potential use of computer aided design techniques by the lay public. The technique is demonstrated in an experiment which involves the builders of a self-build housing association scheme in the design of their own houses.

Abstract: Dynamic ship positioning systems are now required to operate within closer position tolerances and in more adverse weather conditions. There are also stricter requirements on the minimization of energy losses. To optimize performance, recent positioning systems have been designed using Kalman filters and optimal control theory and such a scheme is reviewed briefly. The characteristic locus design method has not previously been used for the design of a stochastic system and it is therefore of interest to see whether or not it may be employed in the above problem. It is shown that not only may this frequency domain design method be used for such systems, but that it has some advantages with regard to the flexibility of the design process. The advantages of using a Kalman filter together with controllers designed by this approach are also demonstrated.

Abstract: Meaningful evaluation of the utility and performance of available nonlinear programming techniques for engineering design optimization requires the availability of realistic design models. Such models have been notably lacking in the process design area. In this paper a detailed equation oriented model of a process for the production of ethylene oxide and ethylene glycol is developed to fill this void. The model involves a large number of variables and equality constraints which are by equation sequencing reduced to 22 independent and equality constraints which are by equation sequencing reduced to 22 independent variables, nineteen constraints, as well as bounds. The salient features of such models, their usefulness and limitations are presented. The applicability of available NLP algorithms and inherent difficulties in their application are discussed. Test results are reported with representative codes. GRG codes are found to provide the only practical approach to large scale design optimization.

Abstract: The Interactive Data Base Designer (IDBD) assumes as input a conceptual description of data to be stored in the data base (in terms of a binary data model) and an expected workload in terms of navigations in the conceptual model. Extensive checking of input is performed. The designer has the possibility to restrict the solution space of the design algorithm by prescribing implementation strategies for parts of the binary model. 1. Chalmers University of Technology, S41296 Gothenburg, Sweden 2. University of Stockholm, S-10691 Stockholm, Sweden 8. This work has been supported by the National Swedish Board for Technical Development Proceedings of the Eighth International Conference on Very Large Data Bases Before we can design a data base schema, compatible with some existing Data Base Management System, we have to determine what kind of data it should contain and what kind of work-load, in terms of queries, updates, inserts and deletes it must be able to handle. In order to permit examination of alternative solutions the requirements must be stated in as implementation independent terms as possible. By 'implementation independent' we mean that there have been made no decisions on how to group data items in records, which access techniques to use and how to navigate in a structure of records and sets. Designing a data base is thus only a (relatively small) part of a systems development process. It is preceded by a number of activities the purpose of which is to analyze corporate information needs and to specify the requirements of an information system to be developed. The Interactive Data Base Designer (IDBD) presented in this paper has been developed to be compatible with two kinds of system design (philosophies) approaches. The first kind, the analytical approach, proceeds through development phases like goal and problem analysis activity analysis etc. and arrives at a comprehensive set of requirements specifications. This set also includes a conceptual information model of relevant parts of the enterprise and a set of information requirements [Bub-801. The conceptual information rode1 (CIM) describes and defines relevant phenomena (entities, relations, events, assumptions, inference rules etc.) of the Universe of Discourse (UoD). The CIM models the UoD in an extended time 108 Mexico City, September, 1982 perspective in order to capture dynamic rules and constraints. The next step in this approach is to 'restrict' the CIM (from a time perspective point of view) and to decide what information to store in the data base and how to conceptually navigate in this set of information in order to satisfy stated information requirements (see Gus-82 for a comprehensive exposition of this problem). If the information to be stored in the data base is defined in terms of a binary data model and the conceptual navigations are specified assuming such a model then this is the required input to IDBD. The other approach to data base design is the experimental one. In this case we assume that a 'fast prototype' is developed by the use of the CS4 system [Ber-77A]. CS4 employs a binary data model and is thus compatible with IDBD. Experimental use of the prototype system can provide us with statistics of navigation types and frequencies. It is, of course, also advantageous to use the experimental approach as a complement to the purely analytical one in order to avoid guesswork concerning the requirements and the work-load. The DBTG-schema design algorithm of IDBD has the binary data model and a set of implementation strategies in common with design-aids developed at the University of Michigan [Mit-75, Ber-77B, Pur-791. It differs, however, from them in several important respects the tool is interactive which gives the designer a possibility to monitor the design process comprehensive checking of the consistency of input data is performed (we have empirically found that it is difficult to supply a tool of this kind with correct input the first time; a waste of time and computer resources is the result of optimizing incorrect input) the designer has a possibility to prescribe certain implementation alternatives for parts of the model (or the whole model). This has the following advantages the designer can test his/hers own solution alternatives which may be 'natural' or which have other, preferred, non-quantifyable properties unnormal or *non-sensesolutions can be avoided Proceedings of the Eighth International Conference on Very Large Data Bases + + the tool can be used to augment an existing DB-schema the solution space can be drastically reduced thereby making IDBD a realistic tool also for design of large, complex data base schemata the description of the work-load is practically realistic as navigation in the conceptual binary model can be defined. This paper describes and explains IDBD in terms of running a small sample case. Section 2 describes the input to IDBD the conceptual binary data model and how to describe the work-load in terms of navigating in the model. User interaction, checking of input and how to supply IDBD with design directives is presented in section 3. The design algorithm and the results of performing a design run are given in section 4.

Abstract: Array-field balance-of-system (BOS) and engineering costs must be reduced for PV power systems to be cost effective for grid-connected applications. Therefore, a study was conducted to develop an innovative and integrated structural and electrical array-field design optimized for minimum life-cycle energy cost, to identify a modular Building Block from this design to be used to construct various sizes of PV array fields at minimum cost, and to standardize the design and prepare complete construction specifications and engineering drawings for reduction of costs of site-specific engineering design and installation. The subsystem area investigated include the support structure, foundation, site preparation, PV module wiring, grounding, ligntning protection, and electromagnetic-interference (EMI) suppression. Maximum use of information from other PV system designs and installation methods were incorporated. Over 50 designs were prepared and evaluated for cost and the final array field selected, developed, and incorporated into a standard Building Block design. Results indicated that the new design greatly reduced BOS costs compared to those of previous installations, provided a high degree of reliability and minimum maintenance, required no major capital investment or long-lead time for an automated plant or equipment, and could be used immediately. The low cost of the array field BOS wasmore » determined to be realistic and economically viable.« less

Abstract: Electric utilities traditionally provide a very high degree of reliability to their customers. In these days of ever increasing costs, utilities are trying to determine how much reliability their customers want and how much they are willing to pay. This is an extremely complex question which involves customer value systems as well as good engineering design. To assist the ultimate customer in making that decision, it is essential that utility companies be able to develop cost-effectiveness indices for the various types of service which might satisfy the customer needs. One vital part of any electric system is its substations. The reliability of substations must be high. However, once a reasonable level of reliability is achieved, there must be a means of evaluating the cost of potential changes to the substations to improve their reliability. This study deals primarily with the reliability of the substation based on its configuration. Thus the comparisons are for cost versus reliability of the design (configuration). The FMEA study using the General Electric reliability programs, COMPUTE and FAILURE, can be used to develop reliability or effectiveness indices. These indices combined with cost data result in quantitative cost-effectiveness data that can be used in comparing substation arrangements.

Abstract: We introduce a global design methodology for large-scale real-time systems; it is based on such concepts as data-flow analysis, sequential processes, communication links, abstract architecture. These concepts give us a guide for designing real-time systems; two tools are also introduced, having a specific action in the design process: OGIVE, a Petri net analyzer, makes the verification and the validation of the abstract structure; OSCAR, analytical queuing network, oriented, evaluates quantitatively some implementation choices of the real system. These tools are integrated very early in the design process to make us sure we shall not have dramatic regressions to do for, eventually, redesigning the system.

Abstract: The design of laboratory experiments is a complex and important scientific task. The MOEGEN project has been developing computer systems for automating the design process in the domain of molecular biology. SPEX is a second-generation system which synthesizes the best ideas of two previous MOLGEN hierarchical planning systems: stepwise refinement of skeletal plans and a layered control structure. It has been tested successfully on several problems in the task domain and promises to serve as a testbed for future work in explanation, experiment debugging, and empirical evaluation of different basic design strategies.

Abstract: The IBM Engineering Design System is a corporate-wide electronic design automation system used for the development of the 3081 and other machines incorporating LSI chips and their carriers. Initially planned in the late 1960s, the system has evolved to respond to a changing set of requirements. This paper describes the evolution of the system data base, as well as the steps taken to ensure the integrity of the data and the correctness of the design.

Abstract: Performance modeling and analysis of computer systems are often ignored during the project design phase in favor of other techniques collectively known as structured design or software engineering. We describe benefits that can result from including performance analysis as an integral part of the design process. Several different goals, time frames, and roles played by performance analysis during system design are illustrated by three case studies of current projects at Los Alamos.

Abstract: In recent years, increases in complexity of hardware/firmware systems, and the concern for systems reliability have resulted in growing interest in methodologies and tools for the design, description and verification of computer systems. A vital component of any such design methodology is the language used for representing the design. In the case of particularly complex systems the design process may involve a succession of stages each of which represents the system at a particular level of abstraction. In such situations several languages may be required, each suited to a particular level of abstraction.We present here, one such family of languages, consisting at present of two members, S*A, an architectural description language, and S*, a high level microprogramming language schema. These closely related (“kin”) languages may be used collaboratively for the systematic, top-down development of an architecture, down to the microcode level. The resulting descriptions of the architecture provide, in addition, a complete unified document of the multilevel design process.

Abstract: Performance modeling and analysis of computer systems are often ignored during the project design phase in favor of other techniques collectively known as structured design or software engineering. We describe benefits that can result from including performance analysis as an integral part of the design process. Several different goals, time frames, and roles played by performance analysis during system design are illustrated by three case studies of current projects at Los Alamos.

Abstract: Publisher Summary Large-scale computational simulations of physical systems have become a valuable adjunct to theoretical and experimental research efforts in the physical and chemical sciences. They are essential because the computer allows solution of problems not soluble by other methods. The use of detailed modeling in conjunction with theory and experiment has reduced the cost and increased the effectiveness and predictability of research project management. This chapter discusses a number of the decisions that a modeler must make to optimize his efforts on a reactive flow simulation project. It explains the choices that are available from the perspectives of program efficiency, flexibility, reliability, accuracy, and lead time to results. The chapter focuses on the vectorized computation of reactive flows. There are five uses for modeling in research and development programs in general, and in the reactive flow modeling in particular. These uses are (1) to evaluate concepts, (2) to interpret measurements, (3) to extrapolate parameters, (4) to calibrate understanding, and (5) in engineering design.

Abstract: The design and implementation of two major comparative experiments is reviewed with particular emphasis on testing methodology. These studies took place at Purdue University. The first, an investigation of the merits of general purpose nonlinear programming codes with major funding from the National Science Foundation, was conducted over the period 1973 to 1977. The second, an investigation of the relative merits of various geometric programming strategies and their code implementations with funding from the Office of Naval Research, was conducted over the period 1974 to 1978. The various major decisions associated with such studies are discussed, such as the selection and collection of problems and codes, the nature of data to be collected, evaluation criteria, ranking schemes, presentation and distribution of results, and the technical design of the experiment itself. The statistical implications of the results in light of the experiment design are examined, as are the effects of various experiment parameters Such as number of variables, number of constraints, degree of nonlinearity in the objective and constraints, and starting point placement. Finally, recommendations for future experiments are given.

Abstract: Methodologies for conducting training effectiveness evaluations (TEE) of ongoing Army training were detailed. The design process, consisting of a literature search and formative evaluations, was discussed. The developed TEE system was described.

Abstract: A prototype method for including damage tolerance considerations in the design of offshore structures has been developed. We start with a design which has been developed using existing techniques. A finite element analysis is performed and the implied factors of safety of various members are ascertained. Critical members are removed systematically from the structural model and the structural response for each case is determined. The factors of safety are recalculated for each of these crippled conditions. Based on this information the section modulus of each member or a group of members is increased in an optimal manner. This provides alternate load paths to take the redistributed loads and thus augment the damage tolerance of the structure. The redistribution of material in order to obtain the desired factors of safety is achieved by solving a multiple objective, non-linear optimization problem. In the past, multiple objective optimization problems have been solved in the linear domain only, using goal programming techniques. We believe that this is the first application of non-linear goal programming in this field. The method is new and makes an important contribution towards incorporating damage tolerance considerations in the design process.

Abstract: This work describes an attempt to systematically design a hardware resource protection mechanism when given the requirements of a particular language as a target. The design process is formalized as a structured walk through the multidimensional computer design space towards a hypothetical class of optimal machines. Each step in this walk involves a change in the distribution of work between the compiler and run-time system but no change in the source language semantics. The starting point for this walk is the result of a semantic analysis of the language to be implemented; typically, this produces a very high level machine where the compiler, if any, is trivial. The walk ends when no changes result in a net improvement. This does not guarantee that the result is even locally optimal, since the changes tried depend on the ingenuity and persistence of the designer.This design approach has been used to arrive at a practical, general purpose protection mechanism oriented towards the needs of the Ada language (preliminary version). This architecture was evaluated by comparing it with the PDP-11/45. For the purpose of this comparison, the protection mechanism was incorporated into a partially specified PDP-11 like instruction set. The number of bits making up the processor state and the number of operations involved in address computation were evaluated. On this basis, the result appears to be competitive and worth further investigation.