Instituto Interdisciplinario de Ciencias Básicas (ICB), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Universidad Nacional de Cuyo, Mendoza, Argentina
Auto-generated profile, based on publicly available information
Pablo Javier Vidal
Instituto Interdisciplinario de Ciencias Básicas (ICB), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Universidad Nacional de Cuyo, Mendoza, Argentina
In this paper, we present a novel implementation of a Cellular Genetic Algorithm (cGA) model for a multi-GPU platform using NVIDIA's CUDA technology. This multi-GPU cGA model is compared first against a serial version in CPU and then versus an implementation on a single GPU. We divide the different operations of the cGA into distinct sets of instructions called kernels. Using the multi-GPU platform we observe that the speedup with respect to the CPU version ranges from 8 to 771, while it is similar to that of the GPU, with a little overhead in the multi-GPU case. Our results demonstrate that multi-GPU desktops can serve as cost-effective parallel computing platforms to obtain accurate results in very short time, although they need special considerations in order to improve on regular single GPUs.
Gene selection aims at identifying a -small- subset of informative genes from the initial data to obtain high predictive accuracy for classification in human cancers. Gene selection can be considered as a combinatorial search problem and thus can be conveniently handled with optimization methods. This paper proposes a Memetic Cellular Genetic Algorithm (MCGA) to solve the Feature Selection problem of cancer microarray datasets. Benchmark gene expression datasets, i.e., colon, lymphoma, and leukaemia available in the literature were used for experimentation. MCGA is compared with other well-known metaheuristic' strategies. The results demonstrate that our proposal can provide efficient solutions to find a minimal subset of the genes.
9 September 2019 in IEEE Latin America Transactions
The automobile has become an essential element in the daily life of people. In the urban traffic network, an intersection is the ``bottleneck point'' of road network capacity. Most cities in Latin America have a traditional single control system or none. In this context, adequate programming of the traffic light cycles helps to reduce vehicles' waiting time and improve traffic flow. In this paper, we compare five metaheuristics to minimize vehicles' waiting time and maximize the velocity of vehicles in two real scenarios of Argentina and Uruguay. Analysis of experimental results shows that the proposed model optimizes the vehicular flow in a multiple-junction urban traffic network with different topologies and traffic flow.