How to run " intlinprog " with " parallel computing " (whole CPU-performance)

I'm using the function "intlinprog" to solve some MILP-Problems (for example with some binary variables). The function is working well, but MATLAB is only using 1 of my CPU-cores and it takes a lot of time to solve bigger problems at the branch and bound stage! (In my case only 33% of it's real performance at 4 cores in total.)

Therefore I searched through the documentations & web and found at the moment no working solution to use the hole computer performance including "intlinprog" and the "parallel computing" (toolbox).

Using "parpool" or "parpool('local')" OR changing the options like "options = optimoptions('solvername','UseParallel',true);" is not working. (<http://de.mathworks.com/help/optim/ug/using-parallel-computing-with-fmincon-fgoalattain-and-fminimax.html#brjh43b>)

Answer 1

[Alan Weiss]

You are correct. Currently, intlinprog has no way to profit from a parallel pool.

Sorry. We are looking at ways of speeding intlinprog, but I cannot tell you any details until the next version is released.

Alan Weiss

MATLAB mathematical toolbox documentation

[Robert van der kraan]

Hello, it has been two years since your answer and I was wondering if anything has changed. Does Matlab R2018a contains any means to speed up intlinprog (branch and bound stage)? What about parralel computing? Gpu computing?

[Alan Weiss]

The best way to find out what has happened with intlinprog is to read the Release Notes. You will see that there is still no parallel functionality in intlinprog, though the function has been improved in other ways.

Alan Weiss

MATLAB mathematical toolbox documentation