Ensemble space inversion (ENSI) is implemented through the PEST_HP suite (version 18). Using ENSI you can calibrate a complex model quickly. The calibration subspace is comprised of random parameter realisations as well as individual parameters. Realisations can be different for different parameter types. Regularisation seeks a minimum error variance solution within the confines of the working subspace. Fits are good, but over-fitting is avoided.
The model run cost of ENSI is commensurate with that of PESTPP-IES. Meanwhile, subspace usage efficiency is maximized. Performance is good robust and fast, even in highly nonlinear calibration contexts.
Setup for ENSI inversion is easy. You need the same files that you need for PEST_HP or PESTPP-IES, plus one more. This extra file allocates parameters and parameter groups to realisation groups, or to the group of parameters whose values will be estimated individually.
It is early days yet. However use of ENSI suggests that its model run efficiency is greater than that of PEST_HP in highly parameterised cases, and its model-to-measurement fit is better than that of PESTPP-IES.
The only down side of using ENSI is that uncertainty analysis is not automatic. However linear options are available. Nonlinear methods such as DSI and IES obtain a head start if preceded by ENSI calibration.
For those interested, see this video on ensemble space inversion.