Struct NonHomogeneousPoisson 

pub struct NonHomogeneousPoisson {
    pub rate_fn: Box<dyn Fn(f64) -> f64>,
    pub rate_max: f64,
}

Non-homogeneous Poisson process with time-varying rate function lambda(t).

Fields

rate_fn: Box<dyn Fn(f64) -> f64>

Rate function lambda(t).

rate_max: f64

Upper bound on the rate function (for thinning algorithm).

Implementations

impl NonHomogeneousPoisson

pub fn new(rate_fn: Box<dyn Fn(f64) -> f64>, rate_max: f64) -> Self

pub fn arrivals(&self, rng: &mut Rng, duration: f64) -> Vec<f64>

Generate arrival times using the thinning (Lewis-Shedler) algorithm.

1. Generate candidate arrival from homogeneous Poisson(rate_max)
2. Accept with probability lambda(t) / rate_max

pub fn counting_path(&self, rng: &mut Rng, duration: f64) -> Vec<(f64, usize)>

Generate counting process path.

pub fn expected_count(&self, duration: f64, n_points: usize) -> f64

Expected number of arrivals in [0, t] = integral of lambda(s) ds. Computed via simple trapezoidal rule.