ILP Applications in mechanical engineering

• Finite element mesh design (Dolsak et al. 1991-96)
• Electrical discharge machining (Karalic 1995)
• FORS applied to model behavior of human operator controlling two variables:

gap and flow. The rules derived for each variable combined in a model that
a) is comprehensible to the domain expert and
b) directs parameter changes towards the optimal process working region.
• Steel grinding (Karalic 1995)
• FORS applied to learn rules for predicting workpiece roughness

from properties of sound produced during grinding.
• Induced rules involve inequalities between the frequencies of the maximum area peak and the spectrum area central point, as well as linear regression, and

reveal domain properties not found by earlier statistical analysis and propositional
learning methods.