Struct ClarabelSolver 

pub struct ClarabelSolver<DemandId, PortfolioId, ProductId>(/* private fields */);

A solver implementation that uses the Clarabel interior point method for quadratic programming to solve the market clearing problem.

This solver is generally more accurate but can be slower than ADMM-based solvers for large problems. It’s a good choice when high precision is needed.

Implementations

impl<A, B, C> ClarabelSolver<A, B, C>

pub fn new(settings: DefaultSettings<f64>) -> Self

create a new solver with the given settings

Trait Implementations

impl<A, B, C> Default for ClarabelSolver<A, B, C>

fn default() -> Self

Returns the “default value” for a type.

impl<DemandId: Clone + Eq + Hash + Ord + Send + Sync + 'static, PortfolioId: Clone + Eq + Hash + Ord + Send + Sync + 'static, ProductId: Clone + Eq + Hash + Ord + Send + Sync + 'static> Solver<DemandId, PortfolioId, ProductId> for ClarabelSolver<DemandId, PortfolioId, ProductId>

type Error = JoinError

Error type for solver failures

type PortfolioOutcome = PortfolioOutcome

A solver will produce outcomes for a portfolio. This should, at a minimum, include the optimal trade rate. It might also include the effective portfolio price, as well as other aggregated statistics.

type ProductOutcome = ProductOutcome

A solver will produce outcomes for a product. This should, at a minimum, include the clearing price. It might also include the total speed of trade, as well as other aggregated statistics.

type State = ()

A solver may leverage additional data to generate a solution, such as a previous batch’s prices and trades.

async fn solve( &self, demand_curves: Map<DemandId, DemandCurve>, portfolios: Map<PortfolioId, (DemandGroup<DemandId>, ProductGroup<ProductId>)>, _state: Self::State, ) -> Result<(Map<PortfolioId, Self::PortfolioOutcome>, Map<ProductId, Self::ProductOutcome>), Self::Error>

Produce a solution given the batch inputs and the solver state.