Untitled

Linear model decision trees as surrogates in optimization of engineering applications

ML models have beenspecially useful in

The underlying physics

Functional relationships

ℎ(𝑥, 𝑑, 𝑟) = 0

𝑑 = 𝑀𝐿(𝑥)

it can be easily extended to the case of multiple surrogates

Reasons to consider ML models in optimization problems.

solution of the inverse problem

can replace first principles models in optimization problems

can replace black-box type external calculations.

Difference of

linear model decision trees

standart decision

differ

returning lineal progression models rather

than constants at the leaves