Struct RewardCalculator 

pub struct RewardCalculator { /* private fields */ }

Calculator for episode reward scores

Implementations

impl RewardCalculator

pub fn new() -> Self

Create a new reward calculator with default weights

pub fn with_weights(duration_weight: f32, step_count_weight: f32) -> Self

Create a calculator with custom weights

pub fn calculate(&self, episode: &Episode) -> RewardScore

Calculate reward score for an episode

pub fn calculate_adoption_bonus( &self, applied_pattern_ids: &[String], outcome_success: bool, ) -> f32

Calculate adoption bonus for patterns that were recommended AND applied AND succeeded.

This bonus rewards episodes where the agent successfully applied recommended patterns. The bonus is proportional to the number of successfully adopted patterns.

Arguments

• applied_pattern_ids - Pattern IDs that were actually applied
• outcome_success - Whether the episode outcome was successful

Returns

Bonus value between 0.0 and 0.3 (max 30% bonus for 3+ successful adoptions)

Example

use do_memory_core::reward::RewardCalculator;

let calculator = RewardCalculator::new();

// 2 patterns applied successfully
let bonus = calculator.calculate_adoption_bonus(
    &["p1".to_string(), "p2".to_string()],
    true
);
assert!(bonus > 0.0);

// No bonus for failed outcome
let no_bonus = calculator.calculate_adoption_bonus(
    &["p1".to_string()],
    false
);
assert_eq!(no_bonus, 0.0);

Trait Implementations

impl Clone for RewardCalculator

fn clone(&self) -> RewardCalculator

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Default for RewardCalculator

fn default() -> Self

Returns the “default value” for a type.