The model of digital ecosystem of corporate production management in a corporation, carried out by its Owner, is considered. The head of the branch of the corporation (Head) controls the subordinate plant headed by the Director. A corporation's branch output is affected by random external influences. The Head knows exactly the maximum possible volume of outsourcing production at the branch level. But he does not know the maximum possible volume of production (potential) in the plant. The Adviser helps the Head to eliminate this uncertainty. For his part, the Director knows exactly plant’s potential. At the same time, the Owner does not know this potential, nor the possibilities of outsourcing. Thus, the Head can manipulate branch output to influence the decisions of the Owner in order to increase his own promotion. Therefore, the Owner needs to learn how to control the Head in order to maximize branch output. Similarly, the Director can manipulate plant’s output to influence the Head's decisions about his own denomination. Therefore, the Head needs to control the Director in order to increase plant’s output. A complex control mechanism has been found, in which total branch output is maximal. The use of such a mechanism is illustrated by the example of the modernization of freight railway platforms. Prospects for research in this area are associated with the use of more complex machine learning procedures in the corporation's hierarchical management system.