Research on process neural networks based on ant colony algorithm

For improving global convergence ability and training speed,an ant colony process neural network model was proposed.Making use of distributed computing and strong robustness of ant colony algorithm,ant colony algorithm was applied in feedforward process neural network training.Topology structure of ant colony process neural network was given,and discussed on training mechanism of ant colony process neural network,analyzed the calculation features.And ant colony process neural network had been used in the annual GDP forecast of Heilongjiang province,verified the effectiveness of ant colony process neural network.