Struct LbpConvergenceMonitor 

pub struct LbpConvergenceMonitor {
    pub history: Vec<LbpIterStats>,
    pub converged: bool,
    pub diverged: bool,
    pub converged_at: Option<usize>,
}

Convergence monitor for Loopy Belief Propagation.

Records per-iteration residual statistics and detects convergence, divergence, or oscillation (residual increasing after initial decrease).

Fields

history: Vec<LbpIterStats>

Per-iteration stats (most recent first).

converged: bool

Whether the algorithm has converged.

diverged: bool

Whether divergence was detected.

converged_at: Option<usize>

Iteration at which convergence occurred (if any).

Implementations

impl LbpConvergenceMonitor

pub fn new() -> Self

Create a new monitor.

pub fn record(&mut self, stats: LbpIterStats, tolerance: f64)

Record statistics from one iteration.

pub fn is_converged(&self) -> bool

True if the last recorded residual is below tolerance.

pub fn last_residual(&self) -> f64

Return the most recent max residual, or f64::INFINITY if no iterations yet.

Trait Implementations

impl Clone for LbpConvergenceMonitor

fn clone(&self) -> LbpConvergenceMonitor

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LbpConvergenceMonitor

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

Formats the value using the given formatter.

impl Default for LbpConvergenceMonitor

fn default() -> LbpConvergenceMonitor

Returns the “default value” for a type.