DCT Lexicon · Governing Dynamics
Adapted Policies and Nonanticipativity
An adapted policy decides using only what is known at the time of the decision. Nonanticipativity is the mathematical form of "no one can act on tomorrow's news".
DCT Governing Dynamics
In depth
Formal statement
Let be the filtration generated by the observations. A policy is admissible if each is -measurable. In scenario form, decisions must agree on scenarios that are indistinguishable at time .
DCT Governing Dynamics: the element's drawer (/governing-dynamics)
Nonanticipativity duality
In multistage stochastic programming, the nonanticipativity constraints can be dualised; their multipliers are the value of information about the future, and progressive hedging solves the relaxed scenario problems while restoring nonanticipativity in the limit.
DCT Governing Dynamics: the element's drawer (/governing-dynamics)
In DCT
Decision authority in DCT is an information structure: who decides, on which signals, at which gate. Admissibility adds governance to measurability: a policy must also respect the owners and approvals the declaration assigns.
DCT Governing Dynamics: the element's drawer (/governing-dynamics)
Limits
Optimising over non-adapted policies (perfect foresight) overstates value: the gap is exactly the expected value of perfect information, and it is a bias, not an upside.
DCT Governing Dynamics: the element's drawer (/governing-dynamics)
Go next
- Value of Information: The multipliers of nonanticipativity measure the value of information (Pathway 4).
- Belief-State Control: Under partial observation, adaptedness is with respect to beliefs.
- Forward–Backward SDEs and Costates: Adapted controls are the setting of forward–backward SDEs.
DCT Governing Dynamics: the element's drawer (/governing-dynamics)
References
- Rockafellar & Wets (1991), Scenarios and policy aggregation
- Shapiro, Dentcheva & Ruszczyński (2021), Lectures on Stochastic Programming
DCT Governing Dynamics: the element's drawer (/governing-dynamics)
In the book: Vol. I, Ch. 8: Stochastic Enterprise Dynamics · Vol. II, Ch. 6: Optimal Control of Enterprise Systems · Vol. II, Ch. 9: Stochastic Enterprise Optimization
Related: Value of Information · Belief-State Control · Forward–Backward SDEs and Costates
GD.05


