Multi‐Objective Optimisation of Flight Simulation Device Data Package Based on Neural Network Surrogate Models

ABSTRACT To demonstrate compliance of flight simulation device (FSD) data packages with regulatory requirements, certification subjects must meet multiple error tolerance objectives. This paper presents a rapid parameter tuning method based on multi‐objective optimisation utilising neural network surrogate models, which improves the efficiency of multi‐system joint simulation. In comparison to the simulation data before optimisation, the average mean relative error for the longitudinal character