HOPSPACK solves derivative-free optimization problems in a C++ software framework. The framework enables parallel operation using MPI (for distributed machine architectures) and multithreading (for single machines with multiple processors or cores). Optimization problems can be very general: functions can be noisy, nonsmooth, and nonconvex, linear and nonlinear constraints are supported, and variables may be continuous or integer-valued.
|Tags||Scientific Optimization parallel|
|Operating Systems||Windows Linux Mac OS X|
Release Notes: The Mediator initializes citizen ID variables properly. The Conveyor thread no longer spins CPU when some workers become idle. Citizen GSS terminates if it cannot generate a trial point that satisfies linear constraint feasibility tolerances. Citizen subproblem solvers that terminate because the "Objective Target" is met now cause parent solvers to also terminate. Integration with the Sandia National Laboratories DAKOTA toolset was improved. HOPSPACK builds on additional platforms, including Solaris, MinGW, and BlueGene.
Release Notes: This release adds the ability to project a trial point to the nearest point that satisfies all linear constraints. The GSS-MS solver has been modified to accept problems with linear constraints. Support has been fixed for the GCC 4.4.1 compiler. Issues with round-off errors violating constraints have been fixed.