Lemma 4.6.1 — (MV) and (MV=) have the same optimal strategies
ProvedMDPFinance.MeanVariance.mv_eq_mveqA strategy is optimal for (MV) if and only if it is optimal for (MV=): the inequality constraint binds at the optimum, so relaxing it to from changes nothing.
This licenses working with the equality-constrained problem throughout the rest of the section (in particular the Lagrangian of Lemma 4.6.2 is built for the equality version), without losing any optimal strategy of the original inequality-constrained (MV).
Formalization Note. Stated as the Iff the book proves (only the "if (MV) then " direction needs an argument; the converse is immediate since (MV=)'s feasible set is a
subset of (MV)'s with the same objective, and its restriction argument is Proof content, not part
of the statement).
import Mathlib import Definitions.Def_MDPFinance_MeanVariance_MVMarket open MeasureTheory ProbabilityTheory
namespace MDPFinance.MeanVariance
/-- Lemma 4.6.1 (Bäuerle–Rieder, p. 118, PDF 132). A strategy `π*` is optimal for `(MV)` if and
only if `π*` is optimal for `(MV=)`. -/
theorem mv_eq_mveq {Ω : Type*} [MeasurableSpace Ω] {d : ℕ} (M : MVMarket Ω d)
(πstar : ℕ → ℝ → (Fin d → ℝ)) :
M.IsOptimalMV πstar ↔ M.IsOptimalMVeq πstar := by sorry
end MDPFinance.MeanVariance
Confirmed by the mission captain (proposal self-audit).
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.