基于BP神经网络的软件可靠性安全性评价方法

Software Reliability Safety Evaluation Method Based on BP Neural Network

摘要: 本文对FPGA软件可靠性安全性评价技术进行研究，建立了基于FPGA软件开发全生命周期的度量模型。将该FPGA软件可靠性安全性度量模型与BP神经网络相结合，建立了基于BP神经网络的FPGA软件可靠性安全性评价网络。通过航空航天项目对其评价结果进行验证，评价结果与专家评价结果基本一致。与基于层次分析法的传统权重评价方法相比，神经网络评价方法具有更好的准确性、适应性和实用性。

Abstract: With FPGA becoming increasingly popular for use in the field of aerospace, the importance of FPGA software reliability and safety to safety critical systems is particularly prominent. However, research on FPGA software reliability and safety is quite insufficient. Concerning this issue, it is necessary to carry out research on FPGA software reliability and safety. In this paper, a reliability and safety metric model based on FPGA software lifecycle is created and safety evaluation network based on BP neural network theory is established. Moreover, a reliability and safety evaluation tool is developed with MATLAB GUI. Comparing with traditional weight evaluation method based on Analytic Hierarchy Process (AHP), the neural network evaluation method has better accuracy, adaptability and practicability.