Skip to content

DCT Lexicon · Optimization

Expected value uplift

The average, across simulated paths, of the company's value with the plan minus its value without it, both read on the same paths.

Drafted for the author's approval from Vol. I, Definition 15.4 (the value functional) and Vol. II, Definition 9.1(iv); the wording is AXIOM's current glossary text, offered for approval

In depth

In the book

The declared aggregation JJ of running and terminal readouts that the GEOP optimizes; warranted component by component under the objective discipline (Definition 15.4).

Vol. I, Glossary of Key Terms, p. 377 (Value functional), verbatim

Formal statement

Δ=E[V(plan)−V(no action)]≈1N∑i=1N(Vi(plan)−Vi(no action)),\Delta = \mathbb{E}\big[V(\text{plan}) - V(\text{no action})\big] \approx \frac{1}{N}\sum_{i=1}^{N}\big(V_i(\text{plan}) - V_i(\text{no action})\big), both terms on the same paths ii.

Drafted for the author's approval, on the expected-value objective of Vol. II, Definition 9.1(iv)

Common misreading

A common mistake is to treat the average path as the plan.

Vol. II, LOS 9.1, verbatim

How AXIOM uses it

Optimize maximizes the expected uplift subject to the floors and the appetite; every recommendation states its expected effect on value against the signed plan, read on the same paths.

Drafted for the author's approval

In the book: Vol. I, Ch. 15: Mathematical Formulation of Enterprise Optimization · Vol. II, Ch. 9: Stochastic Enterprise Optimization

Related: Expected value · Common random numbers · General Enterprise Optimization Problem

V1.Def.15.4 · V2.Def.9.1 · V2.Prop.15.2