2023/05/01 by Agostino Martinelli, Martinelli, Agostino
Biochemistry, Genetics and Molecular Biology · Computer Science · #Evolutionary Algorithms and Applications #FOS: Mathematics #Gene Regulatory Network Analysis #Optimization and Control (math.OC) #Protein Structure and Dynamics
paper · pdf · doi:10.48550/arxiv.2305.01060
openalex publication_date 2023/05/01 · openalex created_date 2023/05/04 · openalex updated_date 2026/07/28
Identifiability describes the possibility of determining the values of the unknown parameters that characterize a dynamic system from the knowledge of its inputs and outputs. This paper finds the general analytical condition that fully characterizes this property. The condition can be applied to any system, regardless of its complexity and type of nonlinearity. In the presence of time varying parameters, it is only required that their time dependence be analytical. In addition, its implementation requires no inventiveness from the user as it simply needs to follow the steps of a systematic procedure that only requires to perform the calculation of derivatives and matrix ranks. Time varying parameters are treated as unknown inputs and their identifiability is based on the very recent analytical solution of the unknown input observability problem. Finally, when a parameter is unidentifiable, the paper also provides an analytical method to determine infinitely many values for this parameter that are indistinguishable from its true value. The condition is used to study the identifiability of two nonlinear models in the field of viral dynamics (HIV and Covid-19). In particular, regarding the former, a very popular HIV ODE model is investigated, and the condition allows us to automatically find a new fundamental result that highlights a serious error in the current state of the art.