# convert a function into optimization expressions

2 views (last 30 days)
Soumili Sen on 27 Oct 2020
Edited: Soumili Sen on 27 Oct 2020
Dear ,
I am writting a code on optimization based on problem based approach; ie;
syms x,y
fun=a/b; % where a and b have a large value combination of random complex no of x and y.(it is nonlinear and non convex)
x = optimvar('x');
y = optimvar('y');
prob.Objective = fcn2optimexpr(fun,x,y); -----> showing error in this line
prob.Constraints.cons1 =norm(cr)^2<=10; ------> showing error in this line
How I can proceed from here.
Thank you.

Walter Roberson on 27 Oct 2020
The solvers are not compatible with symbolic work.
The problem-based approach does not permit you to inject your own functions. You would need to use the solver-based approach and use functions that did any necessary symbolic work internally, ensuring that whatever they returned did not have any symbolic values.
Or you could create your fun and use matlabFunction to convert it to a function handle and use the solver based approach with it.

#### 1 Comment

Soumili Sen on 27 Oct 2020
Ok.. thanks for suggestion.
one question more. what will be the code to inject my equation within the function. like
syms x,y
fun=a/b; % where a and b have a large value combination of random complex no of x and y.(it is nonlinear and non convex)
f=matlabFunction(fun);
how I will call 'f' within a function?