Equivalence problems in structural identifiability

This paper presents a reformulation of the identifiability problem, on the basis of equivalence between a model and the process it represents. Following a brief review of the major definitions in the literature, the concept of “near-equivalence” is introduced for both theoretical and real world identification problems. This concept is illustrated with a general block diagram and two examples. It is concluded that the theory of identifiability needs to be reformulated if it is be broadly applicable to the real world situations.