This work discusses how Radial Basis Function (RBF) neural networks can have their free parameters defined by evolutionary algorithms (EAs). For such, it firstly presents an overall view of the problems involved and the different evolutionary approaches used to optimize RBF networks. It also proposes a Memetic (ie. evolutionary algorithms (EAs) augmented with local search) RBF networks (MRBF) that adopts the most sequential training algorithm, where weights are updated after each training pattern is presented to the network, to elite individuals (having best fitness) and the so-called batch training mode to the remaining individuals of the population. Experiments using a benchmark problem are performed and the results achieved, using the proposed EA, are compared to those achieved by other approaches. The proposed techniques are quite general and may also be applied to a large range of learning algorithms.