NonSmooth Optimization (NSO) Software

"Laugh at your problems; everybody else does."

- Anonymous

NSO Software available here

General Nonsmooth Optimization

LMBM	Limited memory bundle method for large-scale nonsmooth, possibly nonconvex optimization by N. Karmitsa (Fortran 77 and mex-driver for MatLab users). See LDGB for Fortran 95 version of LMBM.
D-Bundle	Diagonal bundle solver for general, possible nonconvex, large-scale nonsmooth minimization by N. Karmitsa (Fortran 95).
MPBNGC	Proximal bundle method for nonsmooth possibly nonconvex (multiobjective) minimization by M.M. Mäkelä (Fortran 77). The code includes the constraint handling (bound constraints, linear constraints, and nonlinear/nonsmooth constraints). MPBNGC can also be used (free for academic purposes) via WWW-NIMBUS -system.
QSM	Quasi-secant solver for nonsmooth possibly nonconvex minimization by A. Bagirov and A. Ganjehlou (Fortran 77). The user can employ either analytically calculated or approximated subgradients in his experiments (this can be done automatically by selecting one parameter).
SMDB	Splitting metrics diagonal bundle solver for general, possible nonconvex, large-scale nonsmooth minimization by N. Karmitsa (Fortran 95).

Limited memory bundle method for large-scale nonsmooth, possibly nonconvex optimization by N. Karmitsa (Fortran 77 and mex-driver for MatLab users). See LDGB for Fortran 95 version of LMBM.

D-Bundle

Diagonal bundle solver for general, possible nonconvex, large-scale nonsmooth minimization by N. Karmitsa (Fortran 95).

MPBNGC

Proximal bundle method for nonsmooth possibly nonconvex (multiobjective) minimization by M.M. Mäkelä (Fortran 77). The code includes the constraint handling (bound constraints, linear constraints, and nonlinear/nonsmooth constraints). MPBNGC can also be used (free for academic purposes) via WWW-NIMBUS -system.

QSM

Quasi-secant solver for nonsmooth possibly nonconvex minimization by A. Bagirov and A. Ganjehlou (Fortran 77). The user can employ either analytically calculated or approximated subgradients in his experiments (this can be done automatically by selecting one parameter).

SMDB

Splitting metrics diagonal bundle solver for general, possible nonconvex, large-scale nonsmooth minimization by N. Karmitsa (Fortran 95).

Derivative Free Optimization

LDGB	Limited memory discrete gradient bundle solver for derivative free general, possible nonconvex, nonsmooth minimization by N. Karmitsa (Fortran 95). To apply LDGB, one only needs to compute at every point the value of the objective function. The subgradient will be approximated. You can also use this code as Fortran 95 version of LMBM (due to some implementational facts it might use less subgradient evaluations than the previous version).
DDG-Bundle	Diagonal discrete gradient bundle solver for derivative free general, possible nonconvex, nonsmooth minimization by N. Karmitsa (Fortran 95). To apply DDG-Bundle, one only needs to compute at every point the value of the objective function. The subgradient will be approximated.
DGM	Discrete gradient solver for derivative free optimization by A. Bagirov, B. Karasozen and M. Sezer (Fortran 77). To apply DGM, one only needs to compute at every point the value of the objective function. The subgradient will be approximated.
QSM	Quasi-secant solver with discrete gradients. See QSM above.

LDGB

Limited memory discrete gradient bundle solver for derivative free general, possible nonconvex, nonsmooth minimization by N. Karmitsa (Fortran 95). To apply LDGB, one only needs to compute at every point the value of the objective function. The subgradient will be approximated. You can also use this code as Fortran 95 version of LMBM (due to some implementational facts it might use less subgradient evaluations than the previous version).

DDG-Bundle

Diagonal discrete gradient bundle solver for derivative free general, possible nonconvex, nonsmooth minimization by N. Karmitsa (Fortran 95). To apply DDG-Bundle, one only needs to compute at every point the value of the objective function. The subgradient will be approximated.

DGM

Discrete gradient solver for derivative free optimization by A. Bagirov, B. Karasozen and M. Sezer (Fortran 77). To apply DGM, one only needs to compute at every point the value of the objective function. The subgradient will be approximated.

QSM

Quasi-secant solver with discrete gradients. See QSM above.

DC Programming

AggSub	Aggregate subgradient based solver for nonsmooth DC programming (difference of two convex functions) by K. Joki (Fortran 95) and N. Karmitsa (Python).
BEM-DC	Bundle enrichment method for nonsmooth DC programming by N. Karmitsa, A. Bagirov and S. Taheri (Fortran 2003).
DBDC	Proximal double bundle solver for nonsmooth DC programming by K. Joki (Fortran 95).
PBDC	Proximal bundle solver for nonsmooth DC programming by K. Joki (Fortran 95).
NonsmoothDCA	Solver for nonsmooth DC programming by A. Bagirov (Fortran 77). The solver is an implemantation of well-known DCA algorithm by Le Thi Hoai An and Pham Dinh Tao.
TCM	Truncated codifferential solver for nonsmooth DC programming by A. Bagirov (Fortran 77).

AggSub

Aggregate subgradient based solver for nonsmooth DC programming (difference of two convex functions) by K. Joki (Fortran 95) and N. Karmitsa (Python).

BEM-DC

Bundle enrichment method for nonsmooth DC programming by N. Karmitsa, A. Bagirov and S. Taheri (Fortran 2003).

DBDC

Proximal double bundle solver for nonsmooth DC programming by K. Joki (Fortran 95).

PBDC

Proximal bundle solver for nonsmooth DC programming by K. Joki (Fortran 95).

NonsmoothDCA

Solver for nonsmooth DC programming by A. Bagirov (Fortran 77). The solver is an implemantation of well-known DCA algorithm by Le Thi Hoai An and Pham Dinh Tao.

TCM

Truncated codifferential solver for nonsmooth DC programming by A. Bagirov (Fortran 77).

Multiobjective Nonsmooth Optimization

MPBNGC	Multiobjective proximal bundle solver. See MPBNGC above.
MDBDC	Multiobjective double bundle method for nonsmooth DC programming by K. Joki and O. Montonen (Fortran 95). MDBDC is able to handle problems which objective and constraint functions can be presented as a difference of two convex (DC) functions.

MPBNGC

Multiobjective proximal bundle solver. See MPBNGC above.

MDBDC

Multiobjective double bundle method for nonsmooth DC programming by K. Joki and O. Montonen (Fortran 95). MDBDC is able to handle problems which objective and constraint functions can be presented as a difference of two convex (DC) functions.

Links to some other NSO solvers and softwares

Proximal bundle method for nonsmooth DC programming Matlab implementations of solvers for nonsmooth DC programming by W. de Oliveira.
MinNS Solver for nonsmooth (possibly) constrained problems by S. Bochkanov. The solver is part of nonlinear optimization suite in ALGLIB (numerical analysis library). It uses internally the gradient sampling algorithm (C++/C#).
OSGA MatLab package for solving large-scale structured convex optimization by M. Ahookhosh.
SolvOpt Solver for local nonlinear optimization problems is an implementation of Shor's r-algorithm by A. Kuntsevich and F. Kappel. The constraints may be taken into account by the method of exact penalization (MatLab, C and Fortran).
GANSO Programming library for Global And Non-Smooth Optimization by CIAO (C/C++).
GradSamp Gradient sampling solver by J. Burke, A. Lewis, and M. Overton (MatLab).
HANSO Hybrid Algorithm for Non-Smooth Optimization by J. Burke, A. Lewis, and M. Overton (MatLab).
OBOE Oracle Based Optimization Engine for convex minimization by J.-P. Vial and N. Sawhney (C++).
PNEW Bundle-Newton method for unconstrained and linearly constrained NSO by L. Luksan and J. Vlcek (Fortran 77).
PVAR Variable metric bundle method for unconstrained and linearly constrained NSO by L. Luksan and J. Vlcek (Fortran 77).
PBUN Proximal bundle method for unconstrained and linearly constrained NSO by L. Luksan and J. Vlcek (Fortran 77).
PMIN solver for MinMax-problems by L. Luksan (Fortran 77).

Nonsmooth Test Problems

Test problems for large-scale unconstrained, bound constrained and generally constrained NSO by N.Karmitsa (Fortran77 and Fortran95).
Test problems for small-scale unconstrained and linearly constrained NSO by L. Luksan (Fortran77).

Solver-o-matic

Solver-o-matic is an online decision tree for choosing a NSO solver. Solver-o-matic will tell you which method/solver is the most suitable for solving your problem.