Controlling Interstate Conflict using Neuro-fuzzy Modeling and Genetic Algorithms

The paper introduces neuro-fuzzy modeling to the problem of controlling interstate conflict. It is shown that a neuro-fuzzy model achieves a prediction accuracy similar to Bayesian trained neural networks. It is further illustrated that a neuro-fuzzy model can be used in a genetic algorithm (GA) based control scheme to avoid 100% of the detected conflict cases. The neuro-fuzzy model is then suggested as a more suitable option to neural networks as the model offers information transparency in the form of fuzzy rules, as compared to the weights of the neural network