A machine learning approach for optimal disassembly planning

Abstract

With the vast amounts of environmental waste being created on a daily basis, many companies are trying to find ways optimally to reuse and recycle obsolete products. Owing to tedious and intensive nature of optimal disassembly planning, expert systems which ease the decision making process are becoming much more prevalent. This paper discusses one such system where a machine learning approach based on a disassembly Petri net (DPN) and a hybrid Bayesian network (HBN) is used. In particular, this method models the disassembly process and predicts the outcome of each disassembly action by examining the probabilistic relationships between the different aspects of the disassembly process. An overall view of the disassembly process and a simple, specific case are provided to illustrate the operation of this expert system.

Keywords:

• Bayesian network
• optimal disassembly planning
• Petri nets
• machine learning
• expert system

Acknowledgements

This work is supported in part by the Charles A. and Ann Morrow Lindberg Foundation and the Christian R. and Mary F. Lindback Foundation.