MODELER : a case based approach for integrating experience in model formulation systems /
| Main Author: | |
|---|---|
| Other Authors: | , |
| Format: | Thesis Book |
| Language: | English |
| Published: |
1993.
|
| Subjects: | |
| Online Access: | ProQuest, Abstract Link to OAKTrust copy |
| Abstract: | Model formulation is pivotal to any effort in the area of intelligent decision support system (IDSS). The different approaches used in model formulation cover a spectrum that includes: those that emphasize selection of suitable problem or tool representation; those that support the view that model representation and model execution are integrated; and those that treat model formulation separately. Most of these approaches have used techniques such as: logic modelling, knowledge-based representations, graphics, and executable modeling languages. Recent trends in model formulation are starting to focus on a case-based approach. In this approach, previously used model formulation plans are stored in memory as cases. In attempting new problems, cases are recalled in an attempt to apply them as is, or by modifying them for the given situation. This methodology of using cases to solve problems is called case-based planning. The research focuses on the model formulation process in the production planning domain. In an attempt to understand the role of experience in the formulation process of expert modelers, a cognitive model is developed. The basis for this cognitive model is a set of verbalizations obtained from experts' while solving problems in the production planning domain. A protocol analysis conducted on these experts' verbalizations provided evidence of a case based approach that is done opportunistically. Using this cognitive model, design considerations are drawn up for constructing MODELER, a case based system which incorporates experience to support the model formulation process. The MODELER integrates techniques from the blackboard paradigm and case based reasoning approach. The blackboard provides a natural environment for the implementation of an opportunistic formulation process. Case based reasoning provided insights for the recall and storage aspects of experience in the model formulation process. An experiment was conducted with subjects whose experience and exposure to linear programming models was wide ranging. The experiment focussed on the recall aspects of subjects' while performing a simple formulation task. The experiment showed that the subjects' exposure and experience with the problem type does determine the cues/stimuli retrieved from the situation to assist the recall process. |
|---|---|
| Item Description: | Vita. "Major subject: Business Analysis." |
| Physical Description: | xi, 167 leaves : illustrations ; 28 cm |
| Bibliography: | Includes bibliographical references. |