Опыт настройки программ оптимизационной инверсии при малом объеме априорной информации

Before doing an optimization inversion one needs to identify two main types of stochastic interrelations. The first type of interrelations shows the possibility to estimate a vector X (the required properties) by a vector Y (logging data burdened with errors). The second type allows to take into account additional equation linking the components of the vector X. The authors describe an original approach to identify parameters of interrelations using data of well tests. They create an algorithm to build a decision rule for bed classification into product-water classes by vector Y. The results of using this approach are illustrated on a specific object