The combination of neural networks and genetic algorithm for fast and flexible wide ing in digital beamforming

The paper presents an approach of using neural networks to apply slow genetic algorithm solutions for real time applications. Genetic algorithms (GAs) are powerful optimization tools which have been applied to solve many complicated problems in a very wide range of areas. However, GA slowness prevent it from being used in real-time systems. A radial basis function neural network (RBFNN) is exploited to approximate the genetic algorithm´s function. This GA-RBFNN approach makes powerful yet slow genetic algorithm solutions possible for real-time problems. As an example, we have successfully applied the proposed approach for fast solution of the GA based wide ing problem in adaptive digital beamforming.