Feat/find optimal retention

[L-M-Sherlock]

Introduction

Function

fn simulate(config: SimulatorConfig, request_retention: f64, seed: Option<u64>) -> f64

It outputs the simulated knowledge retention for given model parameters, request retention and other simulator configuration.

pub fn find_optimal_retention(config: SimulatorConfig) -> f64

It outputs the optimal retention to maximize the simulated knowledge retention.

Structure

pub struct SimulatorConfig {
    w: [f64; 17],
    deck_size: usize,
    learn_span: usize,
    max_cost_perday: f64,
    max_ivl: f64,
    recall_cost: f64,
    forget_cost: f64,
    learn_cost: f64,
}

• w is the weights of FSRS.
• deck_size is the size of deck for simulation.
• learn_span is the number of days for simulation. For example, learn_span=365 means the user has a deadline in one year.
• max_cost_perday is the cost limit for one day's learning (in seconds). For example, max_cost_perday=1800 means the user spends 30 minutes doing reviews every day.
• max_ivl is the maximum interval.
• recall_cost is the cost for a successful review (in seconds).
• forget_cost is the cost for a failed review (in seconds). It includes the time spent in review stage and relearning stage.
• learn_cost is the cost for learning a new cards (in seconds).