Thesis
Using action research to develop a social technical diagnostic expert system for an industrial environment
Southern Cross University, Graduate College of Management
Doctor of Philosophy (PhD), Southern Cross University
2003
Metrics
1 Record Views
Abstract
In modern warfare, the crew of a military vehicle has little room for unscheduled downtime. Thumbing through hundreds of lint-free pages to find the right answer is no longer acceptable as an enemy patrol could be hunting for them. They need a tool that contains key diagnostic information to assist them in solving the technical problem in the shortest possible time. This apparent need had prompted the Singapore Armed Forces to approach a Singapore Engineering Organization (SEO) to develop a diagnostic expert system (DES) that could meet their operational requirements. The SEO is a large military-vehicle producing firm in Singapore. I was engaged by the organisation in 1996 to design and develop a DES.
There were three shortcomings of my initial diagnostic expert system. (1) The lack of an inquiry process had made the task of knowledge acquisition very difficult for new modelers. (2) The shortage of content representation techniques had posed difficulties for modelers to represent the acquired information metaphorically in a DES. (3) The absence of a mechanism for detecting missing content had resulted in inaccuracies in fault isolation. These shortcomings had resulted in logic faults that made the use of DES for accurate diagnosis more difficult in some circumstances. Action research has been used in this study to address these problems.
Three strategies were derived to overcome the logic faults identified in the first version of DES.
The derivation of a four-stage DES inquiry process
The inquiry process is a problem solving process that amalgamates three key features. (1) It adopts four dialectics used in Dick's version of Checkland's Soft Systems Methodology (SSM) mentioned in Dick (2000h). (2) It applies four of Jung's psychological types as decision-making preferences. (3) It uses a number of discrete action research cycles in each stage of the inquiry process. The entire DES inquiry process is repeated until a modeler is able to capture the vehicle information to the best of his or her ability. In this study, the DES inquiry process was used successfully for modeling an armoured recovery vehicle, a military truck and an armoured carrier. It was also used to refine the models of an armoured fighting vehicle constructed prior to this study.
The development of a connective approach in DES
The need to seek out missing content and new environmental faults has led to the development of a connective approach in DES. This approach is developed by Strawson (1992) as an effective way to understand the fundamental structure of human thinking in the field of analytical philosophy. It uses three dimensions, namely, ontology, epistemology and logic. In DES, the first two dimensions, ontology and epistemology, are used to represent the content of a vehicle operation such as starting, driving, stopping or parking a vehicle. The third dimension, logic, makes use of a disconfirming-evidence-seeking mechanism implemented in this study to detect missing content and new environmental faults. The detection is based on the contradiction of logic.
The incorporation of DES as part of a learning process
Faults that have not been modeled are expected to emerge throughout the lifespan of a given vehicle. To overcome this issue, the use of a DES is considered as part of a learning process. This learning process adopts the concepts of single-loop and double-loop learning developed by Argyris and Schon mentioned in Anderson (1997), Korth (2000) and Dick and Dalmau (2000). Single-loop learning is taken up by deployed DES applications. It enables a crew to get over a faulty situation by following the instructions prompted by a DES during troubleshooting. Double-loop learning is undertaken by modelers to make changes to the models. Modelers have to become involved in double-loop learning because diagnostic knowledge requires the presence of human beings as subjects. This on-going learning process enables new faults encountered during the lifespan of a vehicle to be modeled in a given DES.
The use of the above three strategies has resulted a continuously improving process. This process is able to eliminate any logic fault encountered in using a DES for fault finding. In the long run, the knowledge captured by the process will tum the DES into a holistic expert system.
Details
- Title
- Using action research to develop a social technical diagnostic expert system for an industrial environment
- Creators
- Boon Hou Tay
- Contributors
- Stewart Hase (Supervisor) - Southern Cross UniversityBob Dick (Supervisor) - Southern Cross UniversityShankar Sankaran (Supervisor) - Southern Cross University
- Awarding Institution
- Southern Cross University; Doctor of Philosophy (PhD)
- Theses
- Doctor of Philosophy (PhD), Southern Cross University
- Publisher
- Southern Cross University, Graduate College of Management
- Number of pages
- xxvi, 241
- Identifiers
- 991013330228802368
- Copyright
- © Tay Boon Hou 2003
- Academic Unit
- Graduate School
- Resource Type
- Thesis