Training Neural Networks Using Adaptive Simulated Annealing

Adaptive Simulated Annealing (ASA) is an optimization method and an efficiently as a research strategy and guarantee to find the global optimization solution. This research studies Adaptive Simulated Annealing by applying a set of regular experimenting functions, and measures the efficient of this algorithm from the results that produced from training an artificial neural network. The basic measure is depending on the cost function and the network ability to learning and generalization. Therefore we can know the proprieties of algorithms and its advantages and disadvantages.