2-D Genetic Algorithms for Determining Neural Network Structure and Weights

We propose a two-dimensional genetic algorithm for determining the structure and connection weights for neural networks. This technique defines the network as a two-dimensional chromosome, and performs crossover using two-dimensional substructures. Using problems which partially involve Exclusive-OR logic, or training data that is partly dynamically changing, the proposed two-dimensional algorithm is shown to be more effective than the standard genetic algorithms with one-point, two-point, or uniform crossover in the synthesis of neural networks that have meaningful substructures.