Solvers

A progressive decoupling algorithm for minimizing the difference of convex and weakly convex functions over a linear subspace

• Matlab codes used in the manuscript are available here. The solver requires Gurobi to solve subproblems. For the Two-Stage Stochastic Standard Quadratic optimization, run the file SSQP. For the test problem Two-Stage Stochastic Programming with Decision-Dependent Probability, download the DPME solver from the authors webpage and add our two files CwCPDA.m and TwoSSP.m to DPME's folder.  Run TwoSSP. 

Scenario decomposition with alternating projections

• Matlab codes used in the manuscript "Risk-averse stochastic programs and distributionally robust

• optimization via operator splitting" are available here.

Sequential DC Programming

• Matlab codes of some algorithms given in the manuscript Sequential DC Programming are available here.

Animations of some methods for nonsmooth optimization: pdf and Matlab codes

1. Cutting-plane Method pdf goCPM

2. Proximal Bundle Method pdf goPBM

3. Level Bunde Method pdf goLBM

4. Doubly Stabilized Bundle Method pdf goDSBM

5. Extended Cutting-plane Method pdf goECPM

6. Extended Level Bundle Method pdf goELBM

7. All files

Approximate DC Decomposition Algorithm

• Matlab implementation of the algorithms given in the manuscript "Nonsmooth nonconvex optimization via approximate DC decompositions". The main function is "Roda", and the full package is available here.

Proximal bundle algorithm for DC programs 

• Matlab implementation of the algorithms given in the manuscript "Proximal bundle methods for nonsmooth DC programming". The package with solver, 14 test problems, and the main function Main.m is available here. The package uses the Matlab function quadprog for solving the master programs. This is not the best choice: the numerical experiments in the above report were obtained by employing Gurobi to solve the master programs.

Extended Level bundle Methods 

Below some supplementary results on the paper "Regularized optimization methods for convex MINLP problems".

• Matlab sources and three mixed-binary stochastic problems: SLP-A, SLP-B and SLP-C are available here.

• The Matlab file for generating problem instances of the considered convex MINLP with probability constraints is GenDataCCP.m. The blackbox for this problem is oracleCCP.m, which requires the Matlab implicit function mvncdf.

• Supplementary results are available here.

Optimization with Copulae constraints 

This is a set of Matlab routines I jointly wrote with W. van Ackooij. Numerical results are presented in our joint work "Convexity and optimization with Copulae constraints". The package contains:

• Matlab source for generating problem instances of the considered chance-constrained problems

• 14 Archemedian copulae (as well as they gradients) were implemented 

• 4 Solvers (being two supporting-hyperplane methods) are available

Click here to download the package.

Note: The provided oracle (black-box) solves a Conic programming problem through Gurobi. Solver 1 is supposed to work without  supplementary softwares