The MOEA Framework is a Java library for developing and experimenting with multiobjective evolutionary algorithms (MOEAs) and other general-purpose optimization algorithms and metaheuristics. A number of algorithms are provided out-of-the-box, including NSGA-II, ε-MOEA, GDE3, and MOEA/D.
NLopt is a library for nonlinear optimization that allows one to select from a wide variety of optimization algorithms by changing a single parameter. Its features include both local and global optimization, unconstrained, bound-constrained, or nonlinear-inequality constrained problems, and optimization using function values only or using derivatives if they are available. It was initially begun as a wrapper around several existing optimization packages, but it now also includes original implementations of several algorithms for which no free code was available. It provides interfaces callable from C/C++, Fortran, Matlab, GNU Octave, Python, and GNU Guile.
ulatencyd is a scriptable daemon which constantly optimizes the Linux kernel for best user experience. The default configuration tries reduce the latency for a typical desktop system and protects the system from malicious processes and groups. With a different configuration, all other types of systems can be adjusted as well.
Orc is a library and set of tools for compiling and executing very simple programs that operate on arrays of data. The "language" is a generic assembly language that represents many of the features available in SIMD architectures, including saturated addition and subtraction, and many arithmetic operations.
JCGO (pronounced as "j-c-go") translates (converts) programs written in Java into platform-independent C code that can be compiled (by third-party tools) into highly-optimized native code for the target platform. JCGO is a powerful solution that enables your desktop, server-side, embedded, mobile, and wireless Java applications to take full advantage of the underlying hardware. In addition, JCGO makes your programs, when compiled to native code, as hard to reverse engineer as if they were written in C/C++. The JCGO translator uses some optimization algorithms that allow, together with optimizations performed by a C compiler, the resulting executable code to reach better performance compared with the traditional Java implementations (based on the Just-In-Time technology). The produced executable does not contain nor require a Java Virtual Machine to execute, so its resource requirements are smaller than that required by a typical Java VM. This also simplifies the process of deployment and distribution of an application.
AutoDiff.NET is a pure .NET library that allows a developer to easily compose functions symbolically and then automatically calculates the function's value and gradient at any given point. It can be very useful in conjunction with a gradients-based optimization library. It has been tested to work on Mono 2.10 on Linux and on .NET4 on Windows.
ca-ga is a toy artificial life simulation that uses genetic algorithms on large cellular automata. It uses simple but easily extended DNA that is 8k long by default, though you can take the size out to anything you have time to evolve. It sits under each cell of a 128x128 board and orders operations to transfer energy in the hopes of achieving a kill and breed. The simulation features a mutating fitness function, emergent sex, and a proof of concept real world fitness function. After enough generations, the cells or genes could achieve collectivism and organismhood, coordinating the values of the hotspots that determine board temperature in order to maintain a desired equilibrium. But maybe not. If you work in a fitness function, an optimizing problem solver results.
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.