Parameter estimation for optimal disassembly planning

Due to the rapid increase of technological waste in recent years, it has become necessary to find ways of handling the waste in an economically sound and environmentally benign manner. In order to do so, many groups are attempting to disassemble obsolete products in order to reuse or recycle the various components and/or materials such products are comprised. To ease in the disassembly procedure of these products, this paper describes an expert system, consisting of a disassembly Petri network (DPN) and a hybrid Bayesian network (HBN), for optimal disassembly planning. The DPN, the HBN, the interaction between the DPN and the HBN, and the use of inference with the HBN are briefly discussed. Specifically, the paper focuses on ascertaining the parameters of the nodes in the HBN with the use of data collected during the disassembly process and a method for the structure learning of nodes which influence the logistic nodes in the HBN.