Research on Cost Estimation of Launch Vehicle Based on Grey Neural Network

A cost estimation method for launch vehicles is proposed combining the concepts of machine learning, aims to provide assistance in strategic decision-making processes pertaining to satellite launch activities. First, the characteristics of existing methods for estimating the cost of launch vehicles are analyzed, and draws out the machine learning methods based on the characteristics of the current development of launch vehicles in China. Next, a model algorithm based on a dynamic neural network and grey relational analysis is introduced. This algorithm simplifies the network structure by iteratively eliminating low correlation coefficient nodes, effectively addressing the issue of overfitting in small sample data. Finally, the proposed method is validated through a case study about prediction of the Long March series launch vehicle, demonstrating its feasibility and effectiveness.