Xj. Li et Sj. Tang, GENERAL NECESSARY CONDITIONS FOR PARTIALLY OBSERVED OPTIMAL STOCHASTIC CONTROLS, Journal of Applied Probability, 32(4), 1995, pp. 1118-1137
The partially observed control problem is considered for stochastic pr
ocesses with control entering into the diffusion and the observation.
The maximum principle is proved for the partially observable optimal c
ontrol. A pure probabilistic approach is used, and the adjoint process
es are characterized as solutions of related backward stochastic diffe
rential equations in finite-dimensional spaces. Most of the derivation
is identified with that of the completely observable case.