Forecasting approach of subway tunnels deposition based on ANP-SOM model

The main influencing factors on its nature of complexity and indetermination are analysed deeply in this paper. Super Decisions software of ANP (Network Analytic Hierarchy Process) is used to solve the weights of all the influence factors. Use them as the initial weights to train SOM neural network for the fine-tuning; a new ANP-SOM model which consider to various factors, levels and impact of each other feedback synthetically is pretended. This model is well-adjusted, convergence and precision. It was used in forecasting of Xi´an Tunnel deposition and has achieved a good predicted result.