LAUSR

Urban computing brings powerful computational techniques to bear on such urban challenges as pollution, energy consumption, and traffic congestion. After decades of rapid development, artificial neural networks (ANN) have been successfully applied in many disciplines and have enabled many remarkable research achievements. However, no quantitative method has yet been found that can identify every parameter to optimize a neural network. The BP neural network is most frequently used but suffers from the following defects with respect to complex and multidimensional training data or setting of different parameters, i.e., overfitting, slow convergence speed, trapping in local optima and poor prediction effect, and these obstacles have greatly restricted its practical applications. Therefore, this paper proposes a method that uses ant colony optimization (ACO) to train the parameters and structure of the neural network, optimizes its weight and threshold to solve its defects, and applies the model in the optimization scheme of urban operation and management to verify its effect. The experimental simulation proves that the method in this paper is effective and that it makes certain improvements in the local and global search ability, speed, and accuracy of the neural network.