Abstract:

To cope with the uncertain coupling mechanism and complicated hazard factors association of materiel system accidents, a metrics method of system safety based on accident scenario division & coupling is provided. The hazard factors of the materiel system are divided into four variables ，i.e. hazardous fault, human error，fatal environment and hazardous attribute. Then, the materiel safety metrics is equal to the nonlinear fitting of all the hazard variables. On the basis of this, a new quantum harmony search echo state network (QHS-ESN) model and its algorithm are proposed for the battleplan training safety metrics in subject of “high velocity at low altitude”, in which the quantum harmony search (QHS) is merged together with the echo state network (ESN).The proposed model yields a superior optimization performance over the QHS. Compared with the models of original ESN and harmony search back propagation (HS-BP), further test is done in the nonlinear fitting of battleplan accident scenario hazard variables. And simulation result shows that it has a better effect on materiel safety metrics, taking fitting accuracy and stability into account simultaneously.