Struct InfluenceDiagram 

pub struct InfluenceDiagram { /* private fields */ }

Influence diagram for decision-making under uncertainty.

Example

use tensorlogic_quantrs_hooks::{InfluenceDiagram, NodeType};
use scirs2_core::ndarray::{ArrayD, IxDyn};

let mut id = InfluenceDiagram::new();

// Add chance node (weather)
id.add_chance_node("weather".to_string(), 2, vec![]);

// Add decision node (umbrella)
id.add_decision_node("umbrella".to_string(), 2, vec!["weather".to_string()]);

// Add utility node
id.add_utility_node("comfort".to_string(), vec!["weather".to_string(), "umbrella".to_string()]);

Implementations

impl InfluenceDiagram

pub fn new() -> Self

Create a new empty influence diagram.

pub fn add_chance_node( &mut self, name: String, cardinality: usize, parents: Vec<String>, ) -> &mut Self

Add a chance node (random variable).

pub fn add_decision_node( &mut self, name: String, cardinality: usize, parents: Vec<String>, ) -> &mut Self

Add a decision node.

pub fn add_utility_node( &mut self, name: String, parents: Vec<String>, ) -> &mut Self

Add a utility node.

pub fn set_cpt(&mut self, node: &str, cpt: ArrayD<f64>) -> Result<&mut Self>

Set the conditional probability table for a chance node.

pub fn set_utility( &mut self, node: &str, utility: ArrayD<f64>, ) -> Result<&mut Self>

Set the utility table for a utility node.

pub fn set_decision_order(&mut self, order: Vec<String>) -> &mut Self

Set the decision order explicitly.

pub fn chance_nodes(&self) -> Vec<&InfluenceNode>

Get all chance nodes.

pub fn decision_nodes(&self) -> Vec<&InfluenceNode>

Get all decision nodes.

pub fn utility_nodes(&self) -> Vec<&InfluenceNode>

Get all utility nodes.

pub fn get_node(&self, name: &str) -> Option<&InfluenceNode>

Get a node by name.

pub fn to_factor_graph(&self) -> Result<FactorGraph>

Convert to a factor graph for inference.

Decision nodes are treated as having uniform distributions.

pub fn expected_utility(&self, policy: &HashMap<String, usize>) -> Result<f64>

Compute expected utility for a given policy.

A policy maps decision nodes to their chosen values.

pub fn optimal_policy(&self) -> Result<(HashMap<String, usize>, f64)>

Find the optimal policy that maximizes expected utility.

Uses exhaustive search over all possible policies.

pub fn value_of_perfect_information(&self, node: &str) -> Result<f64>

Compute the value of perfect information for a chance node.

VPI measures how much the expected utility would increase if we could observe the node before making decisions.

pub fn information_parents(&self, decision: &str) -> Vec<String>

Get the information parents of a decision node.

These are the nodes whose values are known when making this decision.

pub fn is_well_formed(&self) -> bool

Check if the influence diagram is well-formed.

A well-formed ID satisfies:

• No cycles
• Decisions have a valid temporal order
• Utility nodes have no children

pub fn num_nodes(&self) -> usize

Get total number of nodes.

pub fn num_decisions(&self) -> usize

Get number of decision nodes.

pub fn num_utilities(&self) -> usize

Get number of utility nodes.

Trait Implementations

impl Clone for InfluenceDiagram

fn clone(&self) -> InfluenceDiagram

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for InfluenceDiagram

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for InfluenceDiagram

fn default() -> Self

Returns the “default value” for a type.