NOMAD - A Blackbox Optimization Software

NOMAD is a C++ implementation of the Mesh Adaptive Direct Search algorithm (MADS), designed for difficult blackbox optimization problems. These problems occur when the functions defining the objective and constraints are the result of costly computer simulations.

Charles Audet talks about blackbox optimization


  • Blackbox, nonsmooth optimization
  • Nonlinear constrained optimization
  • Single or biobjective problems
  • Derivative-free optimization
  • Continuous, integer & categorical variables
  • MADS algorithm
  • Matlab and Python interfaces
  • Parallelism with MPI
  • LGPL license
  • Designed for real problems


Comments, feedback and bug reports:

Version history

2024-01 4.4.0 DiscoMads for managing hidden constraints and handlind discontinuities
2022-12 4.3.1 DMulti-Mads for optimization problems
2022-02 4.2.0 Poll direction ORTHO N+1 QUAD, Java-Swig interface
2021-07 4.1.0 VNS Search, static surrogate, poll direction ORTHO N+1 NEG, Windows
2021-04 4.0.2 New architecture
2018-06 3.9 Nelder-Mead search
2015-08 3.8 Granular variables, surrogate library, robust optimization, Python interface
2015-03 3.7 Anisotropic scaling, shared object libraries, performance improvements
2014-09 3.6 Matlab interface, evaluations in blocks

How to cite NOMAD [bibtex]

References for the theory and algorithms [bibtex]