A heuristic approach to process modeling in systems analysis

The systems development life cycle (SDLC) includes five phases: systems analysis, design, implementation, testing, and maintenance. In the systems analysis phase, data modeling and process modeling are the two major tasks involved. These two tasks are very knowledge- and experience-intensive, thus requiring the expertise of skilled professionals. As a result, there have been many knowledge-based systems for data modeling in the systems development literature. Yet, little attention has been paid to process modeling even though knowledge based approaches seem to be best fit for process modeling. In this thesis, to aid and facilitate the process modeling task, a knowledge-based heuristic approach for process modeling is proposed. A heuristic model is developed through protocol analysis. The heuristic model is characterized in terms of generic tasks and their structures. The model consists of three major tasks: Goal Analysis, Process Analysis, and Goal and Process Interaction Analysis. These three major tasks are further decomposed into smaller generic tasks. To validate the heuristic model the method of computer simulation is used. Process Consulting System (PCS) is a computer implementation of the heuristic model. It is shown that the behavior of PCS effectively captures and covers the general features of the expert\u27s behavior in terms of generic tasks identification, heuristic applications to generic tasks, and control knowledge applications. In turn, it can be concluded that the heuristic model explains the expert\u27s behavior in process modeling of systems analysis sufficiently and correctly. There are two major contributions from this study. First, the resultant knowledge-based system, called Process Consulting System (PCS), can help increase the quality of work and productivity of systems analysts. Second, from the perspective of knowledge engineering, the heuristics model proposed in this study will help achieve great increases in the rates and extent of knowledge generation and diffusion within the systems development community. By explicitly specifying a computational model of process modeling in terms of generic tasks and their structure, it becomes easier to reshape the computational model and/or add newly identified tasks to it