FMS scheduling using goal-directed conceptual aggregation

Presents an integrated knowledge-based approach to scheduling flexible manufacturing systems (FMS) using machine learning and simulation. A new learning heuristic based on conceptual clustering is developed, termed 'goal-directed conceptual aggregation' (GDCA). GDCA differs from other learning heuristics in that it can effectively deal with complex dynamic situations through hierarchical structuring of objectives. Its application to FMS scheduling yields improved overall performance through alleviation of many of the problems faced by traditional scheduling techniques. The authors discuss an implementation of a complex FMS as a simulation model that interfaces with a GDCA-based scheduler.< >