This interface brings the ACADO Integrators and algorithms for direct optimal control, model predictive control and parameter estimation to MATLAB. It uses the ACADO Toolkit C++ code base and implements a thin layer to communicate with this code base.

Key Features

Three available standard interfaces:

  • Use the stand-alone Runge-Kutta and BDF's integrators in MATLAB

  • Compose your optimization problem in a MATLAB environment with familiar MATLAB syntax using the generic optimal control interface

  • Write your own C++ code as a MEX-function and compile it using ACADO for MATLAB build-in MEX-compiler

ACADO code generation from MATLAB:

  • SIMexport: generate efficient, customized Runge-Kutta integrators which can be accessed from MATLAB

  • OCPexport: generate efficient, customized OCP solvers which can be accessed from MATLAB

Link your models to ACADO:

  • Link MATLAB ODE or DAE models

  • Link C++ ODE or DAE models

  • Link C ODE or DAE models with ACADO generated code

  • Provide optional Jacobians for faster calculations

Getting started with the MATLAB interface

To install and use the MATLAB interface you need to have a recent MATLAB version and a C++ compiler installed. Follow these steps to get you started in a few minutes.

Step 1 - Installing a compiler

I'm a Linux / Mac user

Make sure you have installed a recent version of the GCC compiler (at least version 4.4).

I'm a Windows user

Make sure you have installed a recent version the Microsoft Visual C++ compiler, at least version 2010.

Step 2 - Configuring MATLAB

Once a compiler is installed it needs to be linked to MATLAB. Open MATLAB (a recent version of MATLAB is required) and run in command window:

>> mex -setup;

MATLAB returns:

Please choose your compiler for building external interface (MEX) files:
Would you like mex to locate installed compilers [y]/n?

Type “y” and hit enter.

I'm a LINUX / Mac user

MATLAB shows you a list of installed compilers. Enter the number corresponding to the GCC compiler (in this case 1) and hit enter.

The options files available for mex are:

1: /software/matlab/20XX/bin/ :
   Template Options file for building gcc MEX-files

2: /software/matlab/20XX/bin/ :
   Template Options file for building MEX-files via [...]

0: Exit with no changes

Enter the number of the compiler (0-2):

I'm a windows user

MATLAB shows you a list of installed compilers. Enter the number corresponding to the Visual C++ compiler (in this case 2) and hit enter.

Select a compiler:
[1] Lcc-win32 C X.Y.Z in C:\PROGRA~1\MATLAB\R20XX\sys\lcc
[2] Microsoft Visual C++ 20XX [...] in C:\Program Files...

[0] None


Confirm the result by writing “y” and hitting enter:

Please verify your choices:

Compiler: Microsoft Visual C++ 20XX [...]
Location: C:\Program Files\Microsoft Visual Studio X.Y

Are these correct [y]/n?

Step 3 - Building the ACADO interface

Please download the toolkit code. Our suggestion is to always clone the stable branch:

  • Linux and OS X users can check-out the code using GIT directly from a terminal. Please execute the following command in a terminal:

git clone -b stable ACADOtoolkit
  • Windows users can use e.g. TortoiseGit:

    • Open your file manager and make a folder called “ACADOtoolkit”.

    • Open tortoise git application and fill in the fields as described here:

      • Enter the URL for cloning. Our suggestion is to always check out stable branch

      • Point “Directory” field to the ACADOtoolkit folder.

      • Check the “Branch” field and enter stable.

    • Click “OK” and wait for the code to be downloaded.

If for any reason you cannot download the code using GIT or you do not want to use GIT (this is not encouraged!), you can download the code in a zip archive

  1. current code from the stable branch or

  2. current code from the master branch

Those archives are automatically updated after each successfully compiled and tested commit we push to the GIT repository.

Please note you do not need to build ACADO at this stage, you just need to download it. We will refer to the main ACADO folder (ACADOtoolkit) as <ACADO_ROOT>. Open Matlab in this directory.

Navigate to the MATLAB installation directory by running:

cd interfaces/matlab/;

You are now ready to compile the ACADO interface. This compilation will take several minutes but needs to be done only once. Run “make” in your command window:

make clean all;

You will see:

Making ACADO...

and after a while when the compilation is finished:

ACADO successfully compiled.
Needed to compile XYZ file(s).

If you need to restart Matlab, run this make file again
to set all paths or run savepath in your console to
save the current search path for future sessions.

ACADO has now been compiled. As the text indicated every time you restart MATLAB you need to run “make” again to set all paths. When running “make” again no new files need to be compiled and the process will only take a few seconds. However, it is easier to save your paths for future Matlab session. Do so by running “savepath” in your command window (this step is optional).


Step 4 - Running your first example

We will now run the OCP getting started example:

cd examples/ocp/getting_started/

The file getting_started.m contains the ACADO syntax to setup and execute a simple Optimal Control Problem. Run “getting_started” in your terminal to test the execution:


You should see a report similar to the following one:


 1:  KKT tolerance = 2.016e-001     objective value = 6.4478e-001
 2:  KKT tolerance = 2.074e+000     objective value = 4.3516e-001
 3:  KKT tolerance = 1.484e-001     objective value = -2.3787e+000
 4:  KKT tolerance = 9.130e-002     objective value = -2.3441e+000
 5:  KKT tolerance = 1.035e-001     objective value = -2.4338e+000
 6:  KKT tolerance = 5.587e-002     objective value = -2.5326e+000
 7:  KKT tolerance = 2.741e-002     objective value = -2.5766e+000
 8:  KKT tolerance = 1.839e-002     objective value = -2.5959e+000
 9:  KKT tolerance = 1.543e-002     objective value = -2.6105e+000
10:  KKT tolerance = 1.494e-002     objective value = -2.6258e+000
11:  KKT tolerance = 5.624e-003     objective value = -2.6404e+000
12:  KKT tolerance = 1.584e-004     objective value = -2.6456e+000
13:  KKT tolerance = 1.214e-008     objective value = -2.6456e+000

convergence achieved.

A graph will be drawn with the results which are stored in the variable 'out’. You're done!

Would you like to read more? Download the user manual.