@article{Alves_Oliveira_Andrade_Nogueira_2021, title={GPU-BRKGA: A GPU accelerated library for optimization using the biased random-key genetic algorithm}, volume={20}, url={https://latamt.ieeer9.org/index.php/transactions/article/view/5233}, abstractNote={<p>This paper presents a CUDA/C++ library, called GPU-BRKGA, for accelerating the biased random-key genetic algorithm (BRKGA). Our library features an easy to use API that allows designers with little experience in GPU development to accelerate the solution of their optimization problems. It automatically handles the problem-independent modules of BRKGA, and leaves the designer in charge with only the decoding procedure, that is, to convert a random-key vector into a solution to the optimization problem. We compared our solution with two other BRKGA API implementations: the standard CPU-based BRKGA API implementation (brkgaAPI) and another GPU-based BRKGA API from the literature (brkgaCuda). Experimental results show that GPU-BRKGA is up to 28x faster than brkgaAPI and up to 3x faster than brkgaCuda.</p>}, number={1}, journal={IEEE Latin America Transactions}, author={Alves, Derek and Oliveira, Davi and Andrade, Ermeson and Nogueira, Bruno}, year={2021}, month={Aug.}, pages={14–21} }