Bounded rationality

Bounded rationality is a framework for understanding decisions made with limited information, time, and computational capacity in a particular environment. Instead of assuming a decision-maker can enumerate every option and optimize across every consequence, it studies the procedures that are actually feasible.

Simon developed the underlying critique in his earlier organizational work and formalized influential models in 1955 and 1956. Historical accounts should distinguish that development from claims about the first appearance of the exact phrase, which was consolidated across his later work.

Satisficing is one procedure within the framework: set an aspiration level, search sequentially, and select an option that meets the threshold. Bounded rationality is broader than satisficing and broader than any catalog of cognitive biases.

Bounded rationality changes the object of study. Instead of comparing a real decision-maker only with an ideal optimizer, it asks which search, inference, and stopping procedures are feasible under the actual limits of information, time, and computation.

That perspective matters in economics, organizational theory, cognitive psychology, and artificial intelligence. Simon received the 1978 Nobel Memorial Prize in Economic Sciences for research on decision processes within economic organizations, work for which bounded rationality was central.

The framework does not by itself say which procedure is best. A useful evaluation must name the environment, goal, information costs, comparator, and benchmark. A procedure can be feasible but inaccurate, or non-optimal under a formal model yet effective for the decision-maker’s real constraints.

Simon’s 1947 Administrative Behavior developed the critique of comprehensive administrative rationality. His 1955 paper, “A Behavioral Model of Rational Choice,” formalized decision-making with limited computational resources and aspiration levels.

His 1956 paper, “Rational Choice and the Structure of the Environment,” made the environment the other half of the analysis. A simple procedure cannot be judged without asking what information structure it encounters.

Later programs took different paths. Tversky and Kahneman studied systematic deviations from normative rules in selected tasks. Gigerenzer and colleagues modeled when simple heuristics exploit environmental structure. Both are important, but neither should be treated as the sole definition of Simon’s framework.

Simon: satisficing and aspiration levels

A satisficing procedure sets an aspiration level, searches alternatives sequentially, and stops when an option clears the threshold. It reduces search and computation but does not guarantee a global optimum. Its performance depends on the threshold, search order, and option environment.

Tversky and Kahneman: heuristics and biases

This program studies when judgments based on representativeness, availability, anchoring, framing, or related processes depart systematically from a normative rule. Each effect must be evaluated in its task, with its benchmark and replication record.

Gigerenzer and colleagues: fast-and-frugal heuristics

This program specifies rules such as take-the-best, which orders cues by validity and stops at the first cue that discriminates between two options, and the recognition heuristic, which applies when exactly one of two options is recognized. It then asks when those rules exploit environmental structure and compete with realistic alternatives. Any less-is-more result is conditional on the task, data, comparator, validation method, and loss function.

How the programs relate

They share an interest in feasible judgment under constraints but differ in questions and standards. A heuristic can depart from one normative inference rule and still predict useful outcomes in a particular environment. Neither observation makes every heuristic biased or every simple rule ecologically rational.

Bounded rationality is a theoretical framework, not a single person-level quantity. Researchers operationalize specific procedures or outcomes: search and stopping rules, framing or anchoring responses, risky and intertemporal choices, predictive accuracy, or information and computation costs.

Each task needs its own reliability, validity, benchmark, and population evidence. Results from one paradigm do not establish a general “bounded rationality score,” and a departure from one normative model is not automatically poor performance on every practical criterion.

Ecological-rationality studies compare defined heuristics with specified alternatives on particular datasets, often using held-out prediction. Some simple rules perform competitively in some environments, but results depend on cue ordering, sample size, dataset structure, comparator, and validation design; they are not a repeated universal advantage for simplicity.

Measures such as maximization, need for cognition, or cognitive reflection assess adjacent tendencies or task performance. They do not measure the bounded-rationality framework itself. The related LifeByLogic task is likewise an educational reflection tool for selected judgment patterns, not a direct or diagnostic measure of bounded rationality.

• Versus optimization. Optimization identifies the best option under a specified model; bounded rationality asks what procedure is feasible when information, time, or computation is limited.
• Versus satisficing. Bounded rationality is the framework; satisficing is one search-and-stopping procedure within it.
• Versus a heuristic. A heuristic is a simplified strategy. Its performance depends on the task and environment.
• Versus cognitive bias. A bias is a systematic pattern relative to a stated benchmark. It is a possible outcome, not a synonym for being bounded.
• Versus analysis paralysis. Analysis paralysis is prolonged failure to commit. Satisficing can provide a stopping rule, but it is not a clinical treatment.

An apartment-search stopping rule

A renter may define acceptable thresholds for price, commute, and safety, search sequentially, and stop at the first apartment that clears all three. This illustrates satisficing: it limits search cost without proving that the selected apartment is globally best. Whether the rule works well depends on the thresholds, search order, market, and cost of continued search.

An illustrative triage tree

A triage protocol might ask one ordered question at a time and stop as soon as a branch specifies the next action. This illustrates the search-and-stopping structure of a fast-and-frugal tree; it is not a claim that this hypothetical tree is validated for coronary-care allocation or any other clinical use. A real clinical decision rule requires population-specific validation, calibration, safety review, and monitoring before use.

• The umbrella is broad. “Bounded rationality” can refer to optimization under constraints, procedural models, heuristics, or departures from classical rationality. Claims are clearer when the procedure is named.
• Description and prescription differ. Showing how people decide does not establish how they should decide.
• Benchmarks differ. Bayesian accuracy, expected utility, prediction, regret, speed, transparency, and robustness can rank the same procedure differently.
• Evidence is effect-specific. Replication or criticism of one bias or heuristic does not validate or invalidate the whole framework.
• Culture and expertise matter. Cue validity, goals, learned strategies, and institutions vary across populations and domains.

• Heuristic — a simplified decision strategy.
• Cognitive bias — a systematic pattern relative to a benchmark.
• Decision support system — software using data, models, or knowledge to support semi-structured decisions.
• Decision hygiene — process safeguards for noise and some bias.
• Analysis paralysis — prolonged indecision or over-analysis.
• Decision fatigue — a proposed change after sustained choice or cognitive effort.

The LifeByLogic Cognitive Bias Test presents brief tasks related to selected judgment effects. Its scores follow the tool’s stated rules and are intended for self-reflection; they do not estimate a general bounded-rationality trait, diagnose decision quality, or show which decision procedure a person uses across environments.

The Big Five Snapshot concerns broad self-reported personality dimensions. It should not be used to infer that a person generally relies on intuitive or deliberative procedures, or that a personality score validates a claim about bounded rationality.

The history is stated conservatively: Simon developed the core argument in earlier work, formalized it in the mid-1950s, and consolidated the terminology across later publications.

• Simon (1955), A Behavioral Model of Rational Choice — formal model of choice under constraints.
• Simon (1956), Rational Choice and the Structure of the Environment — emphasizes the joint role of procedures and environments.
• Stanford Encyclopedia of Philosophy, Bounded Rationality — modern scholarly overview and historical context.
• Nobel Prize facts: Herbert A. Simon — official context for his research on organizational decision processes.
• Gigerenzer and Goldstein (1996) — fast-and-frugal models as one later bounded-rationality program.
• Tversky and Kahneman (1974) — the influential heuristics-and-biases program, related to but not identical with Simon’s framework.

Bounded rationality studies decision procedures that operate with limited information, time, and computation in a particular environment. Simon developed the underlying argument in earlier organizational work and formalized influential models in 1955 and 1956. Satisficing is one possible procedure, not the whole framework.

Later heuristics-and-biases and fast-and-frugal programs emphasize different tasks and benchmarks. The practical lesson is conditional: define the procedure, environment, goal, and comparator before deciding whether a simplified strategy is a useful adaptation or a source of error.

This entry is intended as a citable scholarly reference. Choose the format that matches your context. The retrieval date should reflect when you accessed the page, which may differ from the entry's last-reviewed date shown above.

LifeByLogic. (2026). Bounded Rationality: Definition, Satisficing, and Examples. https://lifebylogic.com/glossary/bounded-rationality/

LifeByLogic. "Bounded Rationality: Definition, Satisficing, and Examples." LifeByLogic, 12 July 2026, https://lifebylogic.com/glossary/bounded-rationality/.

LifeByLogic. 2026. "Bounded Rationality: Definition, Satisficing, and Examples." July 12. https://lifebylogic.com/glossary/bounded-rationality/.

@misc{lblboundedrationality2026,
  author = {{LifeByLogic}},
  title = {Bounded Rationality: Definition, Satisficing, and Examples},
  year = {2026},
  month = {jul},
  publisher = {LifeByLogic},
  url = {https://lifebylogic.com/glossary/bounded-rationality/},
  note = {Accessed: 2026-07-12}
}