Struct optimization_engine::core::fbs::FBSOptimizer

pub struct FBSOptimizer<'a, GradientType, ConstraintType, CostType> where
    GradientType: Fn(&[f64], &mut [f64]) -> FunctionCallResult,
    CostType: Fn(&[f64], &mut f64) -> FunctionCallResult,
    ConstraintType: Constraint,  { /* fields omitted */ }

Optimiser using forward-backward splitting iterations (projected gradient)

Note that an FBSOptimizer holds a reference to an instance of FBSEngine which needs to be created externally. A mutable reference to that FBSEgnine is provided to the optimizer.

The FBSEngine is supposed to be updated whenever you need to solve a different optimization problem.

Implementations

impl<'a, GradientType, ConstraintType, CostType> FBSOptimizer<'a, GradientType, ConstraintType, CostType> where
    GradientType: Fn(&[f64], &mut [f64]) -> FunctionCallResult,
    CostType: Fn(&[f64], &mut f64) -> FunctionCallResult,
    ConstraintType: Constraint, 

pub fn new(
    problem: Problem<'a, GradientType, ConstraintType, CostType>,
    cache: &'a mut FBSCache
) -> Self

Constructs a new instance of FBSOptimizer

Arguments

• problem: problem definition
• cache: instance of FBSCache

pub fn with_tolerance(
    self,
    tolerance: f64
) -> FBSOptimizer<'a, GradientType, ConstraintType, CostType>

Sets the tolerance

Panics

The method panics if the specified tolerance is not positive

pub fn with_max_iter(
    self,
    max_iter: usize
) -> FBSOptimizer<'a, GradientType, ConstraintType, CostType>

Sets the maximum number of iterations

pub fn with_max_duration(
    self,
    max_duration: Duration
) -> FBSOptimizer<'a, GradientType, ConstraintType, CostType>

Sets the maximum number of iterations

Trait Implementations

impl<'life, GradientType, ConstraintType, CostType> Optimizer for FBSOptimizer<'life, GradientType, ConstraintType, CostType> where
    GradientType: Fn(&[f64], &mut [f64]) -> FunctionCallResult + 'life,
    CostType: Fn(&[f64], &mut f64) -> FunctionCallResult + 'life,
    ConstraintType: Constraint + 'life, 

fn solve(&mut self, u: &mut [f64]) -> Result<SolverStatus, SolverError>

solves a given problem and updates the initial estimate u with the solution