Struct argmin::solver::trustregion::TrustRegion

pub struct TrustRegion<R, F> { /* private fields */ }

Trust region method

The trust region method approximates the cost function within a certain region around the current point in parameter space. Depending on the quality of this approximation, the region is either expanded or contracted.

The calculation of the actual step length and direction is performed by a method which implements TrustRegionRadius, such as:

• Cauchy point
• Dogleg method
• Steihaug method

Requirements on the optimization problem

The optimization problem is required to implement CostFunction, Gradient and Hessian.

Reference

Jorge Nocedal and Stephen J. Wright (2006). Numerical Optimization. Springer. ISBN 0-387-30303-0.

Implementations

impl<R, F> TrustRegion<R, F>

where F: ArgminFloat,

pub fn new(subproblem: R) -> Self

Construct a new instance of TrustRegion

Example

use argmin::solver::trustregion::{CauchyPoint, TrustRegion};
let cp: CauchyPoint<f64> = CauchyPoint::new();
let tr: TrustRegion<_, f64> = TrustRegion::new(cp);

pub fn with_radius(self, radius: F) -> Result<Self, Error>

Set radius

Defaults to 1.0.

Example

let cp: CauchyPoint<f64> = CauchyPoint::new();
let tr: TrustRegion<_, f64> = TrustRegion::new(cp).with_radius(0.8)?;

pub fn with_max_radius(self, max_radius: F) -> Result<Self, Error>

Set maximum radius

Defaults to 100.0.

Example

let cp: CauchyPoint<f64> = CauchyPoint::new();
let tr: TrustRegion<_, f64> = TrustRegion::new(cp).with_max_radius(1000.0)?;

pub fn with_eta(self, eta: F) -> Result<Self, Error>

Set eta

Must lie in [0, 1/4) and defaults to 0.125.

Example

let cp: CauchyPoint<f64> = CauchyPoint::new();
let tr: TrustRegion<_, f64> = TrustRegion::new(cp).with_eta(0.2)?;

Trait Implementations

impl<R: Clone, F: Clone> Clone for TrustRegion<R, F>

fn clone(&self) -> TrustRegion<R, F>

Returns a copy of the value.

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

Performs copy-assignment from source.

impl<'de, R, F> Deserialize<'de> for TrustRegion<R, F>

where R: Deserialize<'de>, F: Deserialize<'de>,

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl<R, F> Serialize for TrustRegion<R, F>

where R: Serialize, F: Serialize,

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl<O, R, F, P, G, H> Solver<O, IterState<P, G, (), H, (), F>> for TrustRegion<R, F>

where O: CostFunction<Param = P, Output = F> + Gradient<Param = P, Gradient = G> + Hessian<Param = P, Hessian = H>, P: Clone + ArgminL2Norm<F> + ArgminDot<P, F> + ArgminDot<G, F> + ArgminAdd<P, P>, G: Clone, H: Clone + ArgminDot<P, P>, R: Clone + TrustRegionRadius<F> + Solver<O, IterState<P, G, (), H, (), F>>, F: ArgminFloat,

fn name(&self) -> &str

Name of the solver. Mainly used in Observers.

fn init( &mut self, problem: &mut Problem<O>, state: IterState<P, G, (), H, (), F> ) -> Result<(IterState<P, G, (), H, (), F>, Option<KV>), Error>

Initializes the algorithm.

fn next_iter( &mut self, problem: &mut Problem<O>, state: IterState<P, G, (), H, (), F> ) -> Result<(IterState<P, G, (), H, (), F>, Option<KV>), Error>

Computes a single iteration of the algorithm and has access to the optimization problem definition and the internal state of the solver. Returns an updated state and optionally a KV which holds key-value pairs used in Observers.

fn terminate( &mut self, _state: &IterState<P, G, (), H, (), F> ) -> TerminationStatus

Used to implement stopping criteria, in particular criteria which are not covered by (terminate_internal.

fn terminate_internal(&mut self, state: &I) -> TerminationStatus

Checks whether basic termination reasons apply.