Noemi Paz

Maintenance Scheduling: Issues, Results and Research Needs

Productivity, based on estimated and actual hours, of most maintenance workers is only 30 to 50 per cent. Given the significance of maintenance to manufacturing competitiveness, it is surprising how little research is being carried out. Scheduling is a crucial component of maintenance management and is a focus of research. Identifies the areas of concern in maintenance scheduling and surveys representative work from the academic and practitioner literature. Specific points of practice and theory which need further investigation are pinpointed.

Market timing: a test of a charting heuristic

Abstract We implement a graphical (or ‘charting’) heuristic, the ‘bull flag’, which accepts a particular pattern of historical prices as a signal for a future market price increase, test it with several years of New York Stock Exchange Composite Index history, and find positive results. The results support the validity of technical analysis for stock market price prediction and fail to confirm the efficient markets hypothesis.

The development of knowledge for maintenance management using simulation

The management of maintenance is an area of concern for any industry that depends on the smooth running of equipment to produce a product or carry out a mission at profit or low cost. Maintenance managers must have access to advanced information systems to help them plan their work forces and control operating costs efficiently. This paper describes a method and demonstrates its use to develop a knowledge base for a maintenance supervisor assistant system (MSAS). MSAS interacts with the maintenance manager on a periodic basis to select, for the next period of operations, the proper policies and techniques to meet objectives. The first stage of the method is the knowledge acquisition phase. For this phase, an object-oriented computer simulation model has been developed as a testbed for examining different scheduling heuristics and manning policies in a range of maintenance environments. The dimensions of the environment considered include: preventive maintenance policies, staffing policies, downtime costs, simultaneous downtime practices, travel time impacts, and backlog policies. The dependent variables of interest include: overall machine availability, critical machine availability, worker utilization, cost of the maintenance function, and work order completion time. The second stage is a knowledge engineering effort to codify what is learned from the stage one simulation experiments into a knowledge base for a MSAS. A procedure for deriving expert system rules from simulation experiments is demonstrated. This is followed by validation of the knowledge base through re-employment of the simulation model. >

The Development of Educational Replicas of Software Application Systems

This article presents techniques for developing simplified, functional versions of software applications for training purposes. Representatives of this type of educational software are explained, the design rationale and process are reviewed, and implementation methods are outlined. Three short case studies, exemplifying three categories of this replica software, are included.

Progress report: improving the stock price forecasting performance of the bull flag heuristic with genetic algorithms and neural networks

We back-test a pattern-based heuristic from stock market technical analysis on price and volume time series data for Alcoa Aluminum Companys common stock. Promising results are obtained using a pattern matching approach implemented with spreadsheet technology. Improvement in these results are attained through the application of neural networks and genetic algorithms. Results are confirmed statistically.

Graphical representation of help knowledge for intelligent tutoring systems

Abstract Today an expected facility of interactive computer systems and application programs is an extensive, well-designed, and well-written online help system. A help system frequently provides online tutorial text concerning the operation of the system or application. In the case of Intelligent Tutoring Systems (ITS), an embedded help system provides help with topics in the training domain. In this environment, the help system not only provides user initiated help, but also help text introduced by the embedded computer expert for the purpose of directing or correcting student actions. Explanatory material contained in help systems is called help documentation, and is usually accessed through a “HELP” command. Identification of relevant help texts is difficult for a user, and is a special case of the general information retrieval problem. This paper discusses methods for the retrieval of help knowledge items using graphical, on-screen, tree structures. The configuration of these “study trees” conforms to the hierarchical nature of the tutorial information being referenced. Selection of natural language terms affords initial entry to the trees. A user of the help system browses the trees directly on-screen and consults the subject help texts at will. The study tree presentation facilitates effective selection of pertinent help texts and relates the collection of help material into an integrated whole.

Information systems configuration by expert system: the case of maintenance management

Abstract Expert systems are a technique for transferring expertise from one group to another. In cases where there are few experts and many who need their services, expert systems can be cost effective. The development of maintenance management information systems is one such area. There are few experts in systems analysis and design of maintenance management information systems and a great many who need these systems, especially in a country with an aging equipment base and a shortage of capital to renew that base. This paper is a progress note concerning our plans, analysis, and conjectures in this area.