A Model to Coordinate Interests in Investment Management.

The portfolio selection problem is treated as a two-player game: one player [Unit Investment Trust (UIT)] deals with investments, and the other (agent) allocates funds for it. The game is described in formal terms, and its solution is found. Note that the condition of sustainable development is derived directly from the game solution. UIT's learning is analyzed. Three possible statements of the UIT's learning problem are investigated, each corresponding to a particular machine learning model. An online learning model is considered to predict the information necessary for the agent's optimal behavior in the game. A computational experiment is provided: the three learning models are applied, and the simulation results are discussed. [ABSTRACT FROM AUTHOR]