Software Open Access
GAIM is a flexible and lightweight C++ library implementing Genetic Algorithms (GA) and the Island Models (IM) that incorporate them for real-type genomes. For more information about GAs, IMs, and evolutionary computation (EC) in general, see (Eiben and Smith 2003, Engelbrecht 2007, de Jong 2006, or Mitchell 1998). The library can be called from C/C++ (since an ABI is already implemented) and supports multithreading for running the sort of independent experiments needed to establish Monte Carlo average and best-so-far estimates of attained fitness. In addition, GAIM can be integrated with other packages such as neural simulators (CarlSim) or deep learning frameworks (C++ Pytorch API) in order to perform neuroevolution (i.e. to evolve the weights and other model parameters of artificial neural networks).