3. Uncertainties and conflicts in the parameter estimation

Accounting the effects of uncertainty in the fluid-phase equilibria, modeling became important in the recent decade. Generally, the uncertainty tried to be resolved via probability and random process theories. However, the probabilistic methods can lead to the unreliable estimation of parameters. There are three main complimentary models of uncertainty described in literature:


Three models of uncertainty cannot exist one by one in parameter estimation, and mathematical tools should reveal this fact. The conventional single-criterion methods of parameter estimation are examples of lopsided vision of multicriteria decision-making problem where only one set of variables, strongly dependent on decision-maker experience, is recommended. A lack of single meaning of optimality concept, following from the basics of modern game theory, points to impossibility to construct a completely formal algorithm for parameter estimation and futility of eliminating the uncertainty in the search process. It is more correct to consider the trade-off or rational parameters, which adequately describe thermodynamic properties and phase behavior, generated by EoS.

The following sequence of decision-making steps in model parameter estimation is proposed:

