TY - JOUR
T1 - Conditional nonlinear expectations
AU - Bartl, Daniel
N1 - Publisher Copyright:
© 2019
PY - 2020/2
Y1 - 2020/2
N2 - Let Ω be a Polish space with Borel σ-field F and countably generated sub σ-field G⊂F. Denote by L(F) the set of all bounded F-upper semianalytic functions from Ω to the reals and by L(G) the subset of G-upper semianalytic functions. Let E(⋅|G):L(F)→L(G) be a sublinear increasing functional which leaves L(G) invariant. It is shown that there exists a G-analytic set-valued mapping P
G from Ω to the set of probabilities which are concentrated on atoms of G with compact convex values such that E(X|G)(ω)=sup
P∈P
G(ω)
E
P[X] if and only if E(⋅|G) is pointwise continuous from below and continuous from above on the continuous functions. Further, given another sublinear increasing functional E(⋅):L(F)→R which leaves the constants invariant, the tower property E(⋅)=E(E(⋅|G)) is characterized via a pasting property of the representing sets of probabilities, and the importance of analytic functions is explained. Finally, it is characterized when a nonlinear version of Fubini's theorem holds true and when the product of a set of probabilities and a set of kernels is compact.
AB - Let Ω be a Polish space with Borel σ-field F and countably generated sub σ-field G⊂F. Denote by L(F) the set of all bounded F-upper semianalytic functions from Ω to the reals and by L(G) the subset of G-upper semianalytic functions. Let E(⋅|G):L(F)→L(G) be a sublinear increasing functional which leaves L(G) invariant. It is shown that there exists a G-analytic set-valued mapping P
G from Ω to the set of probabilities which are concentrated on atoms of G with compact convex values such that E(X|G)(ω)=sup
P∈P
G(ω)
E
P[X] if and only if E(⋅|G) is pointwise continuous from below and continuous from above on the continuous functions. Further, given another sublinear increasing functional E(⋅):L(F)→R which leaves the constants invariant, the tower property E(⋅)=E(E(⋅|G)) is characterized via a pasting property of the representing sets of probabilities, and the importance of analytic functions is explained. Finally, it is characterized when a nonlinear version of Fubini's theorem holds true and when the product of a set of probabilities and a set of kernels is compact.
KW - Choquet capacity
KW - DISCRETE-TIME
KW - DUALITY
KW - Dual representation
KW - Dynamic programming
KW - MODEL UNCERTAINTY
KW - Mathematical finance under uncertainty
KW - Nonlinear expectations
KW - Tower property
KW - UTILITY MAXIMIZATION
UR - http://www.scopus.com/inward/record.url?scp=85064819584&partnerID=8YFLogxK
U2 - 10.1016/j.spa.2019.03.014
DO - 10.1016/j.spa.2019.03.014
M3 - Article
SN - 0304-4149
VL - 130
SP - 785
EP - 805
JO - Stochastic Processes and their Applications
JF - Stochastic Processes and their Applications
IS - 2
ER -