The inverse problem of dynamically loaded journal bearings was solved using generalized Reynolds equation coupled with a complete mass conservative cavitation boundary conditions, as outlined by the Jacobsson-Floberg and Olsson (JFO) cavitation theory. In the course of solution, the modified Thomas algorithms was employed, instead of standard Gauss±Jordan reduction method, which fully utilizes the sparse character of the system matrix, and thus greatly reduces computational time. The developed model was tested against the well-known mobility method for the case of journal bearings in a commercial reciprocating air compressor. It was found that the mobility method overestimates minimum film thickness and underestimates such parameters as lubricant flow rate and bearing power loss. In general, the level of error is acceptable for most industrial applications. However, for the journal bearing where the feed pressure is time dependent and starvation effects are predominant, the mobility method may produce large not acceptable errors.