Validation & verification of a Bayesian network model for aircraft vulnerability

Abstract

This paper provides a methodology for Validation and Verification (V&V) of a Bayesian Network (BN) model for aircraft vulnerability against Infrared (IR) missile threats. The model considers that the aircraft vulnerability depends both on a missile’s performance as well as the doctrine governing the missile’s launch. The model is a Knowledge Based System (KBS) and therefore has a knowledge base which consists of both expert knowledge and simulated data which acts as input to the model and is used during inferencing to understand how variables interact. A widely accepted process to certify that a model is suitable for use is the Verification, Validation and Accreditation (VV&A) procedure and is followed in this paper. Throughout the V&V procedure, similarities are drawn between this VV&A process and the well-known Vee-model.