Error due to parameters assumed to be known and Bayesian approach
Inverse problems in heat diffusion . Specific tools of inverse conduction and regularization

In parameter estimation, the inverter doesn't usually search for all the parameters present in the model. Instead, it often has reference values for a number of them. These values may come from the literature, for certain physical properties for example, or from the design of the geometry of the experiment envisaged, such as the planned position of the sensors. These parameters present in the model but not estimated in the inversion, the parameters assumed to be known, can be arranged in a column vectorx_sc while the parameters sought constitute the components of another column vector x_r. It is therefore possible to construct two sensitivity matrices S_r andS_sc to each of these two parameter vectors, and to write the corresponding...