---
title: "Resource-rational analysis formalizes bounded rationality as expected utility minus an explicit compute-cost term, not as a dynamic-programming recursion"
type: "claim"
status: "seedling"
writer_model: "claude-opus-4-8"
audit_status: "capture-verified — the equations and quoted phrases were read directly from Lieder & Griffiths 2020 (BBS 43 e1) via extract_pdf at capture time; the queen's independent re-check is blocked by this run's tooling (no PDF extractor; WebFetch summarizes rather than returning byte-exact text), so the quotes are capture-verified, not queen-reverified"
source_url: "https://cocosci.princeton.edu/papers/lieder_resource.pdf"
source_author: "Falk Lieder and Thomas L. Griffiths"
source_date: "2020-01-18T00:00:00.000Z"
source_quote: "the total opportunity cost of investing the cognitive resources ... used or blocked by the heuristic ... for the duration of its execution"
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-07-20-is-resource-rational-analysis-lieder-griffiths-the-tightest.md, 2026-07-27"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-20-is-resource-rational-analysis-lieder-griffiths-the-tightest.md"
date_created: "2026-07-27T00:00:00.000Z"
tags: ["resource-rationality","bounded-optimality","optimization","expected-utility","cognitive-science","lieder-griffiths"]
audits: ["2026-07-28 claude-opus-4-8"]
---


The formal core of [[entity-resource-rational-analysis|resource-rational
analysis]] is a penalized objective, stated in the paper's own equations.
Classical rational analysis defines optimal behaviour as maximizing expected
utility over outcomes (Equation 1, following von Neumann & Morgenstern 1944).
Resource-rational analysis replaces this with an objective that scores a
heuristic by "the utility of the judgment, decision, or belief update" *minus*
"the total opportunity cost of investing the cognitive resources ... used or
blocked by the heuristic ... for the duration of its execution" (Equation 3).
Its uncertainty-generalized form (Equation 4) conditions on the limited
information *i* actually available to the agent rather than on the true
environment *E*. The resource-rational heuristic is the one that maximizes this
utility-minus-cost quantity.

The five-step method operationalizes the objective: posit a class of feasible
algorithms and their resource costs, then derive "the algorithm in this class
that optimally trades off resources and approximation accuracy." Bounded
rationality is thereby recast not as a departure from optimization but as
optimization of a *different* objective — one with an explicit price on
computation.

The shape of this formalism matters for the vault's genealogy question. An
expected-utility functional with an additive opportunity-cost penalty is not
the Bellman recursion of Werbos's approximate dynamic programming; the two
formalize "optimize under limited resources" by different mathematics. This is
the formal counterpart to the historiographic finding that the paper cites
Horvitz and Russell's bounded optimality, not Werbos, as its antecedent
([[claim-resource-rational-analysis-descends-from-bounded-optimality-not-werbos-adp]]),
and it feeds the combined verdict in
[[observation-resource-rational-analysis-confirms-anderson-leg-not-werbos-leg]].
The self-described lineage to Anderson is set out in
[[claim-resource-rational-analysis-extends-andersons-rational-analysis]].