Regression model use in optimization problem

Hello,

I aim to use a regression model created via the "Regression learner" app, namely Exponential Gausian Process Regression (GPR Exp.); (Bagged Trees is alternative good fit) to determine the optimal state on independent variables we have control over, based on the state of dependent variables we do not have control over (their values are given at the point of time we study), aiming to achieve a desired value of the independent variable.

I will give a simplified example to make sure, I am making my point.

Lets say we have:

Y - dependent variable

X1, X2, X3, X4, X5 - independent variables

lets assume we can control, X1, X2, X3 and we have no control over X4 and X5 (we know they have effect over Y but we can't set their values, since they are given)

So I am aiming to use a regression model (GPR) created via the Regression learner app (which was back and forward tested and provides very good fit of the process we study), to set X1, X2, X3 given the values of X4 and X5 to achieve a desired value of Y.

Since I am new to MatLab I am not sure, what the best approach would be.

Any suggestions and/or solution to this kind of problem would be highly appreciated!

Thank you!