Digital Twins

A digital twin is a set of data and relationships between data. The double allows to carry out a "virtual" experiment on the object, which means to form the best trajectory of development and a way of management. The data in this case is a set of properly ordered multidimensional structures in time and characteristics. For example, a physical customer in a bank is characterized by a debit transaction (X1), a loan transaction (X2), the availability of a loan (X3) and necessarily the time axis T.

If the data structures describing the business entity are created incorrectly in the data warehouse, we will not implement a digital double. Most often it turns out that the data model describes disparate characteristics of an entity that are inconsistent in structure and dynamics. In this case, a digital double is impossible, creating and maintaining multiple storefronts becomes a long and expensive occupation, and the data management service grows unreasonably.

There are three types of data connections in a financial organization.

? The financial model is the most general and the main one. It operates with derived indicators from absolute data and contains a lot of manual semantic settings. Some elements of the financial model are mathematically connected, and the built-in mathematical optimizer (simulator) allows you to find the best management trajectory.

? Mathematical relations are applicable for modeling absolute (quantitative, volumetric) indicators. Mathematical models of characteristics X1, X2, X3,..., Xn are part of the financial model.

? The process relationship is not defined by formulas, but describes methods for the flow of some indicators into others using conditional operators and a sequence of transformations. Optimization mathematics is also applied in process communication, which makes it possible to find the most efficient process.