What is Belady's Algorithm?

Belady's Algorithm is a mental model used for better thinking and decision-making.

How do you apply Belady's Algorithm?

To apply Belady's Algorithm, identify situations where this framework is relevant, then use it as a lens to evaluate your options and decisions. The model is most useful when combined with other complementary mental models.

What category does Belady's Algorithm fall under?

Belady's Algorithm falls under the Computer Science & Algorithms category of mental models. Other models in this category can be found on the Computer Science & Algorithms hub page.

Why is Belady's Algorithm important?

Belady's Algorithm is important because it provides a structured way to think about problems that would otherwise be approached with intuition alone. Understanding this model helps you avoid common reasoning errors and make better decisions.

Where does Belady's Algorithm come from?

Belady's Algorithm is discussed in the tradition of Belady.

Belady's Algorithm Mental Model…

Belady's Algorithm Mental Model… | Faster Than Normal

Section 2

How to See It

Systems

You're seeing Belady's Algorithm when a cache or eviction policy is compared to "optimal" (OPT) or when someone says "we can't do better than Belady without a crystal ball." It's the reference standard, not the implementation.

Decisions

You're seeing the same idea when a decision rule is "drop the option we'll need least soon" — e.g. killing the project we'll use last. That's the Belady logic applied to resource allocation when you (roughly) know future demand.

Section 3

How to Use It

Use Belady as the theoretical upper bound: measure your cache or eviction policy against OPT when you have traces. In strategy, the lesson is "evict what you'll need last" — but since the future is unknown, you approximate with heuristics (LRU, LFU, etc.) or with explicit priorities.

Decision filter

"When we must drop something (cache slot, project, feature): are we evicting the thing we'll need soonest or the one we'll need last? We can't know the future, but we can approximate 'needed last' with recency, frequency, or explicit priority."

As a founder

You can't run Belady in real time, but you can use its logic: when cutting scope or reallocating resources, deprioritize or drop what you expect to need last (or least). Use data (usage, revenue, roadmap) to approximate "future use" and make eviction decisions explicit.

Section 5

Founders & Leaders

Jeff BezosFounder & CEO, Amazon, 1994–2021

Amazon's infrastructure and product decisions constantly trade off what to keep and what to drop — capacity, SKUs, features. Bezos emphasised working backwards from customer needs and long-term value. Belady's principle (evict what you'll need last) maps to that: keep what matters most for the future; cut or defer the rest. Founders can apply it when pruning roadmap or reallocating talent.

Section 7

Connected Models

Reinforces

Least Recently Used

LRU approximates Belady by assuming "recently used" correlates with "will be used soon." It's a heuristic for when you don't know the future. Belady is optimal; LRU is the practical stand-in when traces are not available offline.

Tension

Explore-exploit Tradeoff

Belady optimizes for known (or assumed) future use. Explore-exploit says we must sometimes try the unknown. The tension: strict "evict last-needed" can under-explore; use Belady-style eviction for well-understood resources and allow exploration where the future is uncertain.

Leads-to

Optimization Algorithms

Belady is a special case of optimal sequential decision-making with full information. It leads to the broader class: when you have (or approximate) future information, use it to make eviction or allocation choices; when you don't, use robust heuristics.

Section 8

One Key Quote

"Replace the page that will not be needed for the longest time."
— Belady (1966), on optimal page replacement

Section 11

Summary & Further Reading

Belady's algorithm is the optimal cache eviction policy when the future request sequence is known: evict the page needed farthest in the future. Use it as a benchmark; in practice, approximate with LRU or other heuristics, or apply the same "evict what we need last" logic to prioritization and scope cuts.

A Study of Replacement Algorithms for a Virtual-Storage Computer — Belady (1966)

Article

Original formulation of the optimal (MIN) page-replacement algorithm.

Computer Architecture: A Quantitative Approach — Hennessy & Patterson

Book

Standard reference on caches, replacement policies, and how LRU and others compare to OPT.

Algorithms to Live By — Christian & Griffiths (2016)

Book

Accessible treatment of caching, explore-exploit, and when "optimal" policies inform real decisions.

Belady's Algorithm

Continue exploring

Popular Mental Models

Core Idea

How to See It

How to Use It

Founders & Leaders

Connected Models

One Key Quote

Summary & Further Reading

This connects to...

Continue exploring

Popular Mental Models

More like this, in your inbox

Core Idea

How to See It

How to Use It

Founders & Leaders

Connected Models

One Key Quote

Summary & Further Reading

This connects to...