The internal inductance li and the magnetic axis radial position RMAG are important elements in the development of a reliable real-time control system. Specific algorithms have been developed to obtain these quantities and, with the aim of comparing the relative performances, li and RMAG have also been determined with suitable Neural Networks. Both procedures rely on the Shafranov Integrals, which have been computed on the plasma Last Closed Flux Surface using an updated version of the fast boundary code XLOC. The two approaches have been verified on several classes of equilibrium and both in the limiter and x-point phases of the discharges. Their results have been measured against the estimates of the off-line JET equilibrium code EFIT and it has been found that they provide quite good evaluations of the required quantities, since the differences between their estimates and the EFIT ones are of the order of a few percent. They also provide similar performances in terms of accuracy and computational speed and are quite robust in case of poor quality signals. Both methods are consequently suitable for reliable real-time control systems in JET.