libcgrpp: A Python/C library for box-constrained global optimization with continuous GRASP | ||

Ricardo M. A. Silva, Mauricio G. C. Resende, and Panos M. Pardalos | ||

libcgrpp is a library for solving problems of the type: min F(X) subject to L leq X leq U, where X, L, and U are real n-vectors and F is a function that maps points in the real n-dim space to the real line. There is no constraint on F. For example, F can be non-smooth and highly non-linear. libcgrpp was implemented as an embedded Python-in-C code to take advantage of the simplicity offered by the Python programming language in implementing complex multimodal functions. Besides having access to the extensive standard library of Python, any non-standard module or library, such as SymPy can be used to implement a function. An important feature of the library is that the functions implemented in Python are loaded automatically without the need to recompile any code. | ||

Downloads: tech report [PDF: 191k] source code [ cgraspp-0.0.1.tar.gz: 446k ] | ||

Note: libcgrpp can be used freely for research purposes. Commercial use may require a license. |