An expert system for a local planning environment

In this paper, we discuss the design of an Expert System (ES) that supports decision making in a Local Planning System (LPS) environment. The LPS provides the link between a high level factory planning system (rough cut capacity planning and material coordination) and the actual execution of jobs on the shopfloor, by specifying a detailed workplan. It is divided in two hierarchical layers: planning and scheduling. At each level, a set of different algorithms and heuristics is available to anticipate different situations. The Expert System (which is a part of the LPS) supports decision making at each of the two LPS layers by evaluating the planning and scheduling conditions and, based on this evaluation, advising the use of a specific algorithm and evaluating the results of using the proposed algorithm. The Expert System is rule-based while knowledge (structure) and data are separated (which makes the ES more flexible in terms of fine-tuning and adding new knowledge). Knowledge is furthermore separated in algorithmic knowledge and company specific knowledge. In this paper we discuss backgrounds of the expert system in more detail. An evaluation of the Expert system is also presented.