Near-optimum incentive policies in stochastic team problems with discrepancies in goal perceptions

In this paper, we consider a two-agent stochastic team decision problem with a hierarchical decision structure in a general Hilbert space setting. One of the agents has a different perception of the common team objective functional, as quantified in terms of a finite dimensional parameter vector. The other hierarchically superior agent, uninformed about this discrepancy, but endowed with a suitable information structure, designs a near-optimal incentive policy so that the incurred value of the original team functional is arbitrarily close to its global optimum, in spite of the existing discrepancy. The general solution is determined by some orthogonality relations in some appropriately constructed probability measure spaces.