Some specifics of modeling multi-level agent systems with uncertainty using Groves mechanisms

Uncertainty is often named among the most significant factors impeding decision making in systems with partially independent agents, both technical and social. Among such systems, multilevel systems can be distinguished as a hard case, as uncertainty problem there is combined with the problem of agent opportunism. A special class of mechanisms — the Groves mechanisms that satisfy the requirements of truthfulness in dominant strategies and efficiency — has been well studied in the literature. The article shows how, under conditions of multilevel multi-agent systems with uncertainty, the implementation of Groves mechanisms can solve the problem of systematic distortion of the agents behaviour.