Participants will get some practice with a variety of MPS software in this short course.
Impala is optimized and well-suited for simulating categorical variables. It is based on exhaustive multiple point statistics within the training image and an hybrid list-tree approach for an efficient memory usage and an efficient computational time. Impala follows a multiple grids approach, provides proper conditioning and allows for non stationary simulations by taking auxiliary information into account.
DeeSse follows a different strategy consisting in directly sampling the training image to reproduce the spatial statistics of the model. This latter technique is very flexible and offers a broad range of possibilities for simulating categorical and continuous variables and allows for multi-variate simulations.
These MPS algorithms are available in several software packages such as Isatis, Skua-Gocad, Jewel Suite, and Petrel. Within the framework of this course, practical exercises will be conducted with three of them: Isatis, Skua-Gocad, and Petrel.