DOI: 10.21276/sjpms

Pages: 121-126

The acceleration of the urban process poses new challenges to transport, the construction of three - dimensional traffic has become an important way to solve the congestion problem. The subway is an important factor in the appeal of a big city, however, the rapid and efficient mining of digging tunnels becomes the key to cost savings. The purpose of this paper is to optimize the use of excavation machines and maximize the potential of the development of shield machines, thus achieving savings in maintenance time and cost. For the impact of many factors and The existence of non-linear problems between the variables, we propose a radial basis function network with self-learning ability, which can fit the target and the target variable; In view of the problem that the training speed of RBF network learning method is slow, we introduce the genetic algorithm, and we propose a floating-point coding genetic algorithm based on the adaptive mechanism as a learning algorithm for RBF network learning. The final experimental results show that the accuracy of the new algorithm reaches 95.43%, while learning time significantly shortened by about 37%. All those proves that our algorithm is effective for predicting wear amount, which can improve the efficiency of the machine and save a lot of cost, it is valuable to promote.