SOLUTIONS TO SOME STOCHASTIC TEAM PROBLEMS AND ZERO-SUM GAMES WITH NONCLASSICAL INFORMATION ARISING IN COMMUNICATION SYSTEMS.

Summary form only given. Stochastic dynamic optimization problems are said to be of the nonclassical information type if (roughly speaking) not all the relevant information acquired and used at earlier stages is available at future stages. Such problems are generally very difficult to solve, both analytically and numerically, and very few closed-form solutions exist in the literature. The authors treat a subclass of such problems, arising in communications, which admit closed-form solutions. They consider problems with complete statistical description as well as those with incomplete statistical description, with the latter class arising as jamming problems in communications. For these jamming problems the authors seek closed-form saddle-point solutions in pure or mixed strategies.