A knowledge-based system is formulated to guide the search strategy selection process in simulation optimization. This system includes a framework for machine learning which enhances the knowledge base and thereby improves the ability of the system to guide optimizations. Response surfaces (i.e., the response of a simulation model to all possible input combinations) are first classified based on estimates of various surface characteristics. Then heuristics are applied to choose the most appropriate search strategy. As the search is carried out and more information about the surface becomes available, the knowledge-based system reclassifies the response surface and, if appropriate, selects a different search strategy. Periodically the system’s Learner is invoked to upgrade the knowledge base. Specifically, judgments are made to improve the heuristic knowledge (rules) in the knowledge base (i.e., rules are added, modified, or combined). The Learner makes these judgments using information from two sources. The first source is past experience -- all the information generated during previous simulation optimizations. The second source is results of experiments that the Learner performs to test hypotheses regarding rules in the knowledge base. The great benefits of simulation optimization (coupled with the high cost) have highlighted the need for efficient algorithms to guide the selection of search strategies. Earlier work in simulation optimization has led to the development of different search strategies for finding optimal-response-producing input levels. These strategies include response surface methodology, simulated annealing, random search, genetic algorithms, and single-factor search. Depending on the characteristics of the response surface (e.g., presence or absence of local optima, number of inputs, variance), some strategies can be more efficient and effective than others at finding an optimal solution. If the response surface were perfectly characterized, the most appropriate search strategy could, ideally, be immediately selected. However, characterization of the surface itself requires simulation runs. The knowledge-based system formulated here provides an effective approach to guiding search strategy selection in simulation optimization. / Ph. D.

The increasing effort to develop auditing expert systems raises many questions about their impact on public accounting firms. This research examines the status of expert systems in auditing and investigates the possible future impacts of expert systems on auditing firms. The research involved two separate studies. First, a Delphi study involving auditing and expert-systems experts investigated the likelihood of the proposed future impacts of expert systems on auditing firms. The purpose of the Delphi study was not only to identify the most and least likely impacts, but also to explore the reasons why respondents felt they were the most or least likely. The Delphi panel suggests that expert systems will very likely have an impact on auditing firms in the next decade. The most likely impact identified was that use of an expert system for an audit task provides documentation references for audit judgements and reasoning. other specific impacts that were identified as very likely include distribution of expertise, increased ability to handle complex analyses, and improved decision consistency and quality. The panel also indicated that use of expert systems in auditing is very likely to impact the education of auditors. Second, a case study of an auditing firm using an audit planning expert system provided evidence concerning the impact of an expert system in use. The case study confirmed that, even today, expert systems may be used to provide documentation references and enhance decision consistency and quality. In the situation studied, the impacts were most evident for the less experienced users. The primary contribution of the research is to address questions and concerns about the impact of expert systems on the aUditing profession. The pool of potential impacts of expert systems that has been discussed in the literature can now be narrowed to focus on the most likely impacts. This research is the first step in developing a theory of expert systems impacts. It provides (1) the impetus for further research addressing more specific areas of potential expert systems impact and (2) case study evidence about expert systems impacts that are occurring today. Reasons for the most probable impacts of expert systems on auditing in the future are identified. / Ph. D.

The CODER (COmposite Document Expert/Extended/Effective Retrieval) project is a multi-year effort to investigate how best to apply artificial intelligence methods to increase the effectiveness of information retrieval systems. Particular attention is being given to analysis and representation of heterogeneous documents, such as electronic mail digests or messages, which vary widely in style, length, topic, and structure. In order to ensure system adaptability and to allow reconfiguration for controlled experimentation, the project has been designed as a moderated expert system. This thesis covers the design problems involved in providing a unified architecture and knowledge representation scheme for such a system, and the solutions chosen for CODER. An overall object-oriented environment is constructed using a set of message-passing primitives based on a modified Prolog call paradigm. Within this environment is embedded the skeleton of a flexible expert system, where task decomposition is performed in a knowledge-oriented fashion and where subtask managers are implemented as members of a community of experts. A three-level knowledge representation formalism of elementary data types, frames, and relations is provided, and can be used to construct knowledge structures such as terms, meaning structures, and document interpretations. The use of individually tailored specialist experts coupled with standardized blackboard modules for communication and control and external knowledge bases for maintenance of factual world knowledge allows for quick prototyping, incremental development, and flexibility under change. The system as a whole is structured as a set of communicating modules, defined functionally and implemented under UNIX™ using sockets and the TCP/IP protocol for communication. Inferential modules are being coded in MU-Prolog; non-inferential modules are being prototyped in MU-Prolog and will be re-implemented as needed in C++. / M.S.

Safety is a major concern for tunnel operators. Local authorities responsible for tunnel facilities are concerned with developing restrictions for hazardous materials passing through the facility that will reduce the risk of death and injury, to an extent that these restrictions do not burden commerce unnecessarily. Hazardous material regulatory controls for tunnels are extensive, detailed, and subject to constant changes. The general lack of expertise in tunnel personnel and the lack of a scientific basis leading to the development of these regulations have created problems to local tunnel authorities when updating the restrictions, or when faced with new materials introduced by the industry. Traditionally, most regulatory restrictions enforced both at the federal and the local level are based exclusively on subjective estimation by a panel of experts and on political influence. Experts, however, are not readily available and are expensive to maintain. The need for immediate decisions has constrained tunnel operators to rely on their own intuition in addressing real time transport safety problems in tunnel facility. To address some of these problems, this research investigates the application of knowledge engineering tools to develop a consultative regulatory control system. Specifically, this study presents a structural framework for developing a knowledge-based expert system as an aid to decision-making in tunnel transport safety. The regulatory problem is modeled as a classification type of problem, which lends itself neatly to an expert system application. Heuristic problem solver which is commonly used in solving classification problem involves the systematic matching of the attributes of an unknown entity to a set of pre-defined solutions. For this study's application, the regulatory groupings inherent in existing tunnel regulations are the basis for developing the solution space. The computer program developed uses knowledge which specifies the appropriate regulation applicable to a new commodity based on the material's physical and chemical properties. / M.S.

The FOster Care Expert System (FOCES) was developed to provide advice to social workers of the Roanoke City Department of Social Services who must select foster care homes for children who cannot remain with their own families. It was implemented using the General pUrpose Expert Shell System (GUESS) and Horn Clause Prolog. The system's design was greatly influenced by unique features of the problem domain. Among the key concerns were: unresolved questions within the social work profession about foster home selection and evaluation, serious methodological and philosophical difficulties associated with defining a good "person-environment fit", and the volatile, free-form narrative nature of the information maintained by social services agencies about children and homes. "Traditional" approaches to knowledge acquisition and representation adopted by developers of expert systems were of limited use. Adaptation of extended "p-norm" Boolean queries previously used in information retrieval work simplified the knowledge representation and matching tasks for this human services application. Evaluation of FOCES' performance, using a small database of children and homes, has shown that the system can select appropriate foster care placements at least as well as some experienced social workers. / Master of Science

Expert systems are very useful and probably the most fruitful products of applied artificial intelligence. Expert systems, however, are very· expensive to develop. Powerful construction tools are indispensable to construct, modify and maintain a practical expert system. GUESS/l is a domain-independent expert systems shell that captures and enhances the strengths of its predecessors while at the same time overcoming.their limitations. GUESS/l gives a strong emphasis on human engineering, language generality, diversity of data representation and control structures, programming and run-time environment, database construction facilities and security, and many other aspects that are related to the ease of development and maintenance of expert systems. / Master of Science

The purpose of this study is to implement the advantages of the relatively new field of expert systems to robot kinematics. The research presented in this thesis illustrates the progress in combining the two fields. An expert system used to solve the kinematic equations of general purpose robots is presented along with some examples. / Master of Science

Although widely recognized in the research community as one of the most accurate non-destructive methods for the determination of residual stress in polycrystalline structural materials, X-ray diffraction has not been widely adopted in the field. This is partly due to the fact that such measurements require, most often, a well-trained user with knowledge in both materials and mechanical sciences in addition to the specific know-how of the instrument. We believe that computer assistance could contribute to the promotion of this technique by increasing the productivity and accuracy of these measurements. We have developed a prototype of an expert system, using Nexpert Object's shell, to assist a non-trained operator in the validation and interpretation of X-ray diffraction residual stress data. The present work describes this prototype which has been designed to confirm the feasibility of the concept. Its knowledge base contains relevant examples of the rules necessary for data validation. The prototype has also validated most of the concepts required for the implementation of a full-scaled version by evaluating all of the major technical features such as graphics representation, external routine calls, and databases access. We have implemented significant rules to validate an experiment, link our expert system with a database management system, develop a superset of data able to receive output from any existing X-ray machine, and are working with a statistical pattern recognition software to discriminate between various d-vs-sin²Psi curves, to classify our data. / Master of Science

Increased complexity of constructed facilities, owners’ changing needs and international competition have created strong demands for advanced construction technology. Recent developments in computer technologies provide the Architecture-Engineering-Construction (AEC) industry with significant potential for innovation. However, the adoption of state of-the-art computer technologies by a majority of AEC firms faces many hurdles. The objective of this research is the development of methodologies for accelerating the adoption of Expert Systems (ES) technology by the AEC industry. This is accomplished with the development of a Technology Transfer (T²) model for ES. The T² model has been based on the models of innovation developed by Rogers, Shaffer, and Tatum, as well as on case Studies of ES adoption by several organizations. The ES T² model has focused on the following issues: 1) the stages of the T² process and the managerial actions required for successful adoption of ES technology; 2) the economical, technological and organizational factors affecting the T² process; and 3) alternative strategies that managers can deploy to successfully transfer the technology to their firm. The T² model provides a useful framework that can significantly enhance managers’ ability to expedite the transfer of ES technology in the AEC industry. / Master of Science

Scheduling is one of the most important functions in a factory and it is determining when and with what resources jobs should be accomplished. An important factor that affects the scheduling of jobs is the dynamic variation of factory status. Existing computer based scheduling systems do not address the need of making effective decisions dynamically with the variations in factory status. Traditionally, Operations Research techniques have provided an effective tool in solving manufacturing planning problems. But these methods have not been able to effectively address real time control problems in the manufacturing environment. To address some of these problems, this research investigates applying an expert system approach to develop an interactive real time dynamic scheduling system. Specifically, a knowledge base structure is developed and applied to a case study representing a two stage production system. A Blackboard concept has been utilized to organize and maintain the dynamic data base. The major knowledge representation schemes used in the system include, frame structures, relational tables, and production rules. The system was tested on a case study, by conducting a sample interactive session on a set of simulated dynamic situations. The test demonstrated the viability of implementing knowledge based systems for dynamic scheduling at the operational level of a plant. / Master of Science