Struct QualityFunctionMuOracle 

pub struct QualityFunctionMuOracle {
    pub norm_type: NormType,
    pub centrality_type: CentralityType,
    pub balancing_term: BalancingTermType,
    pub max_section_steps: i32,
    pub section_sigma_tol: Number,
    pub section_qf_tol: Number,
    pub sigma_max: Number,
    pub sigma_min: Number,
    pub mu_min: Number,
    pub mu_max: Number,
}

Fields

norm_type: NormType

centrality_type: CentralityType

balancing_term: BalancingTermType

max_section_steps: i32

section_sigma_tol: Number

section_qf_tol: Number

sigma_max: Number

sigma_min: Number

mu_min: Number

mu_max: Number

Implementations

impl QualityFunctionMuOracle

pub fn new() -> Self

pub fn calculate_mu_with_predictor_centering( &mut self, data: &IpoptDataHandle, cq: &IpoptCqHandle, nlp: &Rc<RefCell<dyn IpoptNlp>>, pd_search_dir: &mut PdSearchDirCalc, ) -> Option<Number>

Drive the predictor + centring solves through pd_search_dir, project the results onto the four bound-mask subspaces, then run pick_sigma over a q(σ) closure that builds the σ-step and reduces it to QualityFunctionAggregates before invoking evaluate_quality_function. Mirrors upstream IpQualityFunctionMuOracle.cpp::CalculateMu lines 188-485.

Returns None if either linear solve fails (caller falls back to LOQO, matching upstream’s IpAdaptiveMuUpdate.cpp::CalculateMuFromOracle:330-340).

Trait Implementations

impl Default for QualityFunctionMuOracle

fn default() -> Self

Returns the “default value” for a type.

impl MuOracle for QualityFunctionMuOracle

fn calculate_mu(&mut self) -> Option<Number>

Probe the next mu given the current iterate state. Phase 10 fills in the actual signatures; trait surface here keeps the option matrix matchable in AlgBuilder.