.png){kind=icon}
.png){kind=icon}
Return
Tolerance allocation of complex systems based on supervised machine learning and adaptive sampling
.png){kind=icon}
PRE
AI
DOI：10.1007/s00170-025-17336-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper presents a novel framework for the tolerance allocation of complex systems. The framework integrates adaptive sampling, importance sampling, and the C4.5 decision tree algorithm to enhance exploration of the solution space and improve accuracy. Adaptive sampling dynamically adjusts the sampling strategy based on previous iterations, while importance sampling focuses on regions with a higher probability of yielding conforming solutions. The C4.5 algorithm generates decision trees from labelled data, identifying patterns and rules for conforming configurations, which assists in selecting the most effective candidate solutions for tolerance allocation. When tested on the Janssen mechanism, the framework demonstrated significant improvements in dataset balance and tolerance identification.
Keywords:
Tolerancing
Tolerance allocation
Supervised machine learning
Monte Carlo simulation
Journal
T
IF:
0
Papers:
2.0K
Citations:
0