G-LVQ (Genetic Learning Vector Quantization) is an algorithm that combines features taken from genetic search and neural learning. In this way, it is faster than a genetic algorithm by itself, by adding neural hill-climbing, and samples more efficiently the search space than a neural network alone.
Genetic algorithms alone are usually slow in classification problems, since they are too coarse-grained to obtain a solution quickly. On the other hand, neural algorithms are usually fast (in terms of processor cycles), if they are close to the solution, but tend to get stuck in local minima. Combining both kind of algorithms manages to avoid local minima, and finds solutions accurately.