The APPSPACK project defines a derivative-free local optimizer. This solver is a pattern search method that can optimize continuous applications without the aid of gradients or other derivative information. The APPSPACK solver is a parallel pattern search method that exploits a novel asynchronous parallel search.

The Coliny project includes a variety optimizers that all have a COLIN optimization interface. Coliny defines global optimization methods like multi-start local search, evolutionary algorithms and DIRECT. Additionally, Coliny includes interfaces to other optimizers, including a Lipshitzian solver and the Cobyla optimizer.

The EXACT project defines a software framework for describing, controlling, and analyzing computer experiments. It provides the experimentalist with convenient software tools to ease and organize the entire experimental process, including the description of factors and levels, the design of experiments, the control of experimental runs, the archiving of results, and analysis of results.

The FAST project defines a software framework for coordinating software build tests. It includes hooks for archiving tests in a centralized location, and for generating a simple HTML dashboard. The FAST software can also integrate and summarize results from EXACT experimental tests.

The GNLP project defines a general-purpose global optimization solver that leverages and extends the PEBBL branch-and-bound library. This solver can rigorously solve nonlinear problems that can be expressed algebraicly (e.g. with a mathematical programming language like AMPL).

The OPT++ project defines a library of optimizers for continuous search domains, particularly methods for engineering design.

The ParPCx project defines a parallel linear programming (LP) solver that is based on the serial interior-point code PCx and the Trilinos solver library.

The PEBBL project defines an object-oriented scalable library for branch-and-bound.

The PICO project defines parallel mixed integer linear programming (MILP) solver. Three publicly available linear programming solvers are integrated into Acro for this MILP solver: GLPK, SOPLEX and CLP.

The UTILIB project defines a library of general-purpose C++ utilities. The capabilities of UTILIB are similar in spirit to STL and BOOST.