Influence Diagram¶

A typed directed acyclic graph that joins uncertain variables, choices, available information, and value functions so a decision problem can be evaluated for an expected-utility-maximizing policy.

Version: v2 · 2026-09-19 · History
Domain-specific #: 2060
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomains: Decision Analysis, Graphical Decision Modeling Under Uncertainty → Mathematics
Aliases: Decision Network

Core Idea¶

An influence diagram is a compact formal representation of a decision problem under uncertainty. It is a directed acyclic graph whose node types and incoming-arc semantics jointly distinguish what is uncertain, what can be chosen, what will be known when each choice is made, and what the decision maker values. A fully specified diagram attaches conditional probability models to chance nodes, action sets to decision nodes, and utility or value functions to value nodes. Evaluation then searches for a policy—a rule mapping the information available at each decision to an admissible action—that maximizes expected utility.

How would you explain it like I'm…

The Choosing Map

An influence diagram is a special drawing that helps you make a choice when you don't know everything. It has bubbles for things you can't control, like whether it will rain, and bubbles for things you choose, like taking an umbrella. It also has a bubble for how happy you'll be at the end. Arrows show what you'll know before you choose, so you can pick the plan that's best on average.

Luck, Choice, and Reward Map

An influence diagram is a map of a decision where some things are uncertain. It uses different kinds of boxes: chance boxes for things that happen by luck, decision boxes for things you choose, and value boxes for how good the result is for you. Arrows connect them, and an arrow means something different depending on where it points. An arrow into a chance box says 'this depends on that'. An arrow into a decision box says 'you'll know this before choosing'. An arrow into a value box says 'this affects how good the outcome is'. Once it's filled in, you can work out the best rule for choosing.

Typed Decision Graph

An influence diagram is a formal model of a decision under uncertainty, drawn as a directed graph with no loops. It has three node types: chance nodes (uncertain variables with probabilities), decision nodes (choices with sets of possible actions), and value nodes (utility functions). The meaning of an arrow depends on the node it enters. Into a chance node, it marks a variable that the probability depends on. Into a decision node, it marks information you will observe before choosing. Into a value node, it marks an input to the utility. Solving the diagram means finding a policy, a rule that says what to do given what you know at each decision, that maximizes expected utility. It's like a Bayesian network with choices and preferences added. Compared with a decision tree, it is usually more compact, though a tree can show lopsided scenarios more directly.

An influence diagram is a directed acyclic graph that specifies a decision problem under uncertainty through typed nodes and typed arcs. Chance nodes carry conditional probability distributions, decision nodes carry action sets, and value nodes carry utility functions. Arc semantics depend on the receiving node. A conditional arc into a chance node marks a conditioning parent, an informational arc into a decision node marks a variable observed before the choice, and an arc into a value node marks an argument of the local utility. The graph is therefore a factorized specification of the model, not just an illustration, and these readings should not be lumped together as vague 'influence'. Evaluation seeks a policy that maps each decision's informational parents to an admissible action so as to maximize expected utility. Compared with a Bayesian network, it adds control, information timing, and preference. Compared with a decision tree for the same problem, it usually avoids duplicated subtrees and makes conditional independence explicit, though trees show asymmetric scenarios more directly.

Scope of Application¶

Influence diagrams belong primarily to decision analysis, operations research, and artificial intelligence. They are used when a decision maker must choose actions while outcomes depend on uncertain variables and when the timing of observations matters. Typical work includes model structuring, expert elicitation, policy optimization, sensitivity analysis, and value-of-information analysis. Shachter's direct evaluation algorithm made the representation computational as well as communicative: a well-formed diagram can be transformed by arc reversal and node removal without first enumerating a decision tree.

Clarity¶

The fastest diagnostic is to ask three questions about every arrow: what type of node receives it, what mathematical object does that receiver carry, and what changes if the arrow is removed? If the receiver is a chance node, removal asserts a conditional-independence simplification and changes its probability parents. If it is a decision node, removal withholds an observation from the decision rule. If it is a value node, removal says that local utility does not directly depend on that tail variable.

Manages Complexity¶

A decision tree expands histories. With \(k\) binary observations, even one choice after all observations may require \(2^k\) observation paths, and repeated downstream structures appear on many branches. An influence diagram represents each variable once and records only direct conditioning or information relations. Conditional independence permits the evaluator to reuse local factors rather than enumerate every complete history. This is the central compression: repeated branches become shared variables and local relationships.

Abstract Reasoning¶

The graph licenses several concrete inferences. A missing arc in the chance subgraph can express conditional irrelevance under the graphical semantics, allowing probability factors to be separated. A barren chance node with no path to value and no informational role may be removable. Reversing a suitable chance arc corresponds to a Bayesian refactorization, provided acyclicity and the necessary parent adjustments are preserved. Eliminating chance variables averages their downstream value over the relevant conditional distribution; eliminating a decision maximizes over its alternatives while recording the maximizing rule.

Knowledge Transfer¶

Within its home fields, the same grammar transfers across applications. A clinician choosing a treatment, an engineer choosing whether to inspect a component, and a project manager choosing whether to fund a test can all be represented by chance variables, decision alternatives, observation timing, and value consequences. The variable labels and assessments change; the typed roles and policy criterion remain literal.

Relationships to Other Abstractions¶

Current abstraction Influence Diagram Domain-specific

Parents (1) — outgoing typed relations

Influence Diagram includes Directed Acyclic Graph Prime

instantiated literally.

Hierarchy path (1) — routes to 1 parentless root

Influence Diagram → includes Directed Acyclic Graph → is a kind of Network

Neighborhood in Abstraction Space¶

Influence Diagram sits in a sparse region of the domain-specific corpus (94^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2769 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25