A new tolerance allocation approach based on decision tree and Monte Carlo simulation

The rise of artificial intelligence techniques provides new opportunities in the robust design and tolerancing fields. These technological growths require adapting new practices. In this paper, a new tolerance allocation approach based on supervised machine learning (decision tree) and Monte Carlo simulation is proposed. The most important branch of the decision tree which is generated based on the results of the Monte Carlo simulation, is analysed and the best tolerances to be allocated are defined in terms of quality/cost trade-off. The effectiveness of the proposed approach is demonstrated and discussed through a non-linear problem: a Jansen linkage study.