SCHEDULING IN A GENERIC PERSPECTIVE - KNOWLEDGE-BASED DECISION-SUPPORT BY DOMAIN ANALYSIS AND COGNITIVE TASK-ANALYSIS

Creating computer systems to perform a task that requires expert knowl edge is complex. To a certain extent, the developer has to gain the ta sk expertise himself in order to design adequate decision support. Sev eral methods are available to capture the knowledge of an expert and t o implement this in a knowledge based system. However, these methods d o not make adequate use of previous experiences encountered in develop ment trajectories in the same problem domain. We were confronted with this problem in the domain of planning and scheduling. In building kno wledge based systems for scheduling support (e.g., ZKR, for nurse sche duling) we realized that there could be possibilities for reuse. We sa w several similarities in different scheduling domains as well as in t he systems, but were unable to get a grip on the issue at once. An ade quate apparatus for modeling the scheduling domain, the task execution , and the supporting systems was missing. We developed a framework bas ed on object orientation, cognitive task analysis, and decision suppor t systems. The framework captures the underlying structure of the sche duling task in a modeling approach using ideas from literature about m odeling and reuse such as modeling by analogy, Case Based Reasoning, a nd CommonKads. It can be used to tackle problems in different scheduli ng settings, thereby facilitating the development process of schedulin g support systems.