Our Software libraries
A MATLAB software for nuclear norm regularized least squares problems based on an accelerated proximal gradient method.
A MATLAB software for A Sparse Doubly Nonnegative Relaxation of Polynomial Optimization Problems with Binary, Box and Complementarity Constraints.
(Robust Stochastic Optimization Made Easy) is a MATLAB algebraic toolbox designed for generic optimization modeling under uncertainty.
Based on the robust stochastic optimization (RSO) framework proposed by Chen, Sim, Xiong (2019), RSOME unifies a wide variety of approaches for optimization under uncertainty, including the traditional scenario-tree based stochastic linear optimization, classical robust optimization as well as the emerging distributionally robust optimization that considers state-of-the-art data-driven ambiguity sets.
ySOT (over 88,000 downloads) is an RBF Surrogate Global Optimization Toolbox developed in Python for optimization of expensive black-box multi-modal objective functions with continuous and/or integer variables, single or multiple objectives, w/wo nonlinear constraints, where the number of objective evaluations is limited. Code Reference: https://arxiv.org/pdf/1908.00420.pdf, which lists many related papers.
pySOT is based on papers from Distinguished Professor Shoemaker (ISEM) & co-workers. Prof. Shoemaker’s module ISE6511 (spring ) teaches about pySOT algorithms and other global optimization methods.
pySOT has been applied to science/engineering problems including environmental fluid mechanics, manufacturing systems, hyper-parameters tuning in deep learning, etc.
Github Link: https://github.com/dme65/pySOT; Documentation: https://pysot.readthedocs.io/en/latest/
