SHOT no longer needs an external NLP solver in its primal strategy since SHOT can now call itself for solving fixed NLP problems. Activated with Primal.FixedInteger.Solver=2. For nonconvex problems it is still recommended to use an external NLP solver.
Partition convex nonseparable quadratic functions as separate constraints using an eigenvalue decomposition-based reformulation. Activated with Model.Reformulation.Quadratics.UseEigenValueDecomposition=true.
Support for performing an initial polyhedral approximation of the nonlinear feasible set before feasibility-based bound tightening. Activated with Model.BoundTightening.InitialPOA.Use=true.
Support for problems containing semi-continuous and semi-integer variables.
Support for problems containing special ordered sets.
Minor improvements and bug fixes
Improved support for generating supporting hyperplanes for the entire nonlinear feasible set instead of the feasible sets for the individual constraint functions. Activated with Dual.ESH.Rootsearch.UseMaxFunction=true.
Improved support for passing nonconvex quadratic functions directly to the MIP solver (if supported).
Bug fixes for the AMPL interface.
Open access paper about nonconvex features in SHOT
Date: March 20, 2021
The full paper about the nonconvex features in SHOT is now available online in the Journal of Global Optimization. The paper also includes some benchmarks on nonconvex MINLP and MIQCQP problems.
SHOT is now available as a solver on the NEOS Solver. You can submit jobs in GAMS format here.
The NEOS (Network-Enabled Optimization System) Server is a free internet-based service for solving numerical optimization problems. Visit the NEOS Server web site to access 60 state-of-the-art solvers in more than a dozen categories.
SHOT 1.0 is now available at Github. Binaries are available for Windows, and Linux and MacOS users can easily compile SHOT themselves using the instructions on the page Compiling.
SHOT is also available in GAMS 31.1. A trial version can be downloaded from GAMS.
Preprint available about the nonconvex features in SHOT
Date: March 22, 2020
There is now a new preprint available that describes the nonconvex features in SHOT. The paper also includes some benchmarks on nonconvex MINLP and MIQCQP problems. It can be downloaded from Optimization Online: