MANSO

Multistart Algorithm for Identifying All Optima of Nonconvex Stochastic Functions

MANSO is a multistart algorithm to identify all local minima of a constrained, nonconvex stochastic optimization problem. The algorithm uniformly samples points in the domain and then starts a local stochastic optimization run from any point that is the ``probabilistically best'' point in its neighborhood.

Dependencies

1. Create a python 3.7 environment

    conda create --name py37 python=3.7
    source activate py37
   

2. Install the following Numpy,Scipy,qiskit(if run QAOA), scikit-quant(Snobfit and Bobyqa)

    pip install numpy
    pip install scipy
    pip install qiskit
    python -m pip install scikit-quant
   

3. Require MATLAB engine to use ASTRODF (https://github.com/simopt-admin/simopt/wiki) from Python (<= 3.7)

   (Tested on MATLABR2020a)

   https://www.mathworks.com/help/matlab/matlab_external/install-the-matlab-engine-for-python.html

Run MANSO from the command prompt

1. Help: python MANSO.py --help

2. Example: To find local optima of Branin function (https://www.sfu.ca/~ssurjano/branin.html) use

    python MANSO.py -b 0.1 -t 0.01 -o 0.01 -n 5 -B 15000 -OB 500 -d 2 -P Branin -seed 12 -lso ASTRODF
   

To plot data profiles use Performance_Test.ipynb