Knowledge base maintenance using cultural algorithms: application to the DLMS manufacturing process planning system at Ford Motor Company

Ford´s Direct Labor Management System (DLMS) is a fully-deployed application that is being utilized at Ford´s assembly plants throughout the world. DLMS has been in production since 1990, and maintainability of the knowledge base has become very difficult over time, due to changes in the following areas: the business environment, the processes and physical concepts being modeled, and the underlying hardware and software architecture. We have previously (2000, 2001) described how cultural algorithms can be used in a top-down and bottom-up fashion to re-engineer a dynamic semantic network-based knowledge base. In this paper, we compare the two approaches and discuss how they can be used in tandem to re-engineer a semantic network