Interface InfeasibilityInformationOrBuilder
- All Superinterfaces:
com.google.protobuf.MessageLiteOrBuilder
,com.google.protobuf.MessageOrBuilder
- All Known Implementing Classes:
InfeasibilityInformation
,InfeasibilityInformation.Builder
@Generated
public interface InfeasibilityInformationOrBuilder
extends com.google.protobuf.MessageOrBuilder
-
Method Summary
Modifier and TypeMethodDescriptionType of the point used to compute the InfeasibilityInformation.double
The objective of the linear program labeled (1) in the previous paragraph.double
Let (y_ray, r_ray) be the algorithm's estimate of the dual and reduced cost extreme ray where (y_ray, r_ray) is a vector (satisfying the dual variable constraints) scaled such that its infinity norm is one.double
Let x_ray be the algorithm's estimate of the primal extreme ray where x_ray is a vector that satisfies the sign constraints for a ray, scaled such that its infinity norm is one (the sign constraints are the variable bound constraints, with all finite bounds mapped to zero).double
The value of the linear part of the primal objective (ignoring additive constants) evaluated at x_ray, i.e., c' * x_ray where c is the objective coefficient vector.double
The l_∞ norm of the vector resulting from taking the quadratic matrix from primal objective and multiplying it by the primal variables.boolean
Type of the point used to compute the InfeasibilityInformation.boolean
The objective of the linear program labeled (1) in the previous paragraph.boolean
Let (y_ray, r_ray) be the algorithm's estimate of the dual and reduced cost extreme ray where (y_ray, r_ray) is a vector (satisfying the dual variable constraints) scaled such that its infinity norm is one.boolean
Let x_ray be the algorithm's estimate of the primal extreme ray where x_ray is a vector that satisfies the sign constraints for a ray, scaled such that its infinity norm is one (the sign constraints are the variable bound constraints, with all finite bounds mapped to zero).boolean
The value of the linear part of the primal objective (ignoring additive constants) evaluated at x_ray, i.e., c' * x_ray where c is the objective coefficient vector.boolean
The l_∞ norm of the vector resulting from taking the quadratic matrix from primal objective and multiplying it by the primal variables.Methods inherited from interface com.google.protobuf.MessageLiteOrBuilder
isInitialized
Methods inherited from interface com.google.protobuf.MessageOrBuilder
findInitializationErrors, getAllFields, getDefaultInstanceForType, getDescriptorForType, getField, getInitializationErrorString, getOneofFieldDescriptor, getRepeatedField, getRepeatedFieldCount, getUnknownFields, hasField, hasOneof
-
Method Details
-
hasMaxPrimalRayInfeasibility
boolean hasMaxPrimalRayInfeasibility()Let x_ray be the algorithm's estimate of the primal extreme ray where x_ray is a vector that satisfies the sign constraints for a ray, scaled such that its infinity norm is one (the sign constraints are the variable bound constraints, with all finite bounds mapped to zero). A simple and typical choice of x_ray is x_ray = x / | x |_∞ where x is the current primal iterate projected onto the primal ray sign constraints. For this value compute the maximum absolute error in the primal linear program with the right hand side set to zero.
optional double max_primal_ray_infeasibility = 1;
- Returns:
- Whether the maxPrimalRayInfeasibility field is set.
-
getMaxPrimalRayInfeasibility
double getMaxPrimalRayInfeasibility()Let x_ray be the algorithm's estimate of the primal extreme ray where x_ray is a vector that satisfies the sign constraints for a ray, scaled such that its infinity norm is one (the sign constraints are the variable bound constraints, with all finite bounds mapped to zero). A simple and typical choice of x_ray is x_ray = x / | x |_∞ where x is the current primal iterate projected onto the primal ray sign constraints. For this value compute the maximum absolute error in the primal linear program with the right hand side set to zero.
optional double max_primal_ray_infeasibility = 1;
- Returns:
- The maxPrimalRayInfeasibility.
-
hasPrimalRayLinearObjective
boolean hasPrimalRayLinearObjective()The value of the linear part of the primal objective (ignoring additive constants) evaluated at x_ray, i.e., c' * x_ray where c is the objective coefficient vector.
optional double primal_ray_linear_objective = 2;
- Returns:
- Whether the primalRayLinearObjective field is set.
-
getPrimalRayLinearObjective
double getPrimalRayLinearObjective()The value of the linear part of the primal objective (ignoring additive constants) evaluated at x_ray, i.e., c' * x_ray where c is the objective coefficient vector.
optional double primal_ray_linear_objective = 2;
- Returns:
- The primalRayLinearObjective.
-
hasPrimalRayQuadraticNorm
boolean hasPrimalRayQuadraticNorm()The l_∞ norm of the vector resulting from taking the quadratic matrix from primal objective and multiplying it by the primal variables. For linear programming problems this is zero.
optional double primal_ray_quadratic_norm = 3;
- Returns:
- Whether the primalRayQuadraticNorm field is set.
-
getPrimalRayQuadraticNorm
double getPrimalRayQuadraticNorm()The l_∞ norm of the vector resulting from taking the quadratic matrix from primal objective and multiplying it by the primal variables. For linear programming problems this is zero.
optional double primal_ray_quadratic_norm = 3;
- Returns:
- The primalRayQuadraticNorm.
-
hasMaxDualRayInfeasibility
boolean hasMaxDualRayInfeasibility()Let (y_ray, r_ray) be the algorithm's estimate of the dual and reduced cost extreme ray where (y_ray, r_ray) is a vector (satisfying the dual variable constraints) scaled such that its infinity norm is one. A simple and typical choice of y_ray is (y_ray, r_ray) = (y, r) / max(| y |_∞, | r |_∞) where y is the current dual iterate and r is the current dual reduced costs. Consider the quadratic program we are solving but with the objective (both quadratic and linear terms) set to zero. This forms a linear program (label this linear program (1)) with no objective. Take the dual of (1) and compute the maximum absolute value of the constraint error for (y_ray, r_ray) to obtain the value of max_dual_ray_infeasibility.
optional double max_dual_ray_infeasibility = 4;
- Returns:
- Whether the maxDualRayInfeasibility field is set.
-
getMaxDualRayInfeasibility
double getMaxDualRayInfeasibility()Let (y_ray, r_ray) be the algorithm's estimate of the dual and reduced cost extreme ray where (y_ray, r_ray) is a vector (satisfying the dual variable constraints) scaled such that its infinity norm is one. A simple and typical choice of y_ray is (y_ray, r_ray) = (y, r) / max(| y |_∞, | r |_∞) where y is the current dual iterate and r is the current dual reduced costs. Consider the quadratic program we are solving but with the objective (both quadratic and linear terms) set to zero. This forms a linear program (label this linear program (1)) with no objective. Take the dual of (1) and compute the maximum absolute value of the constraint error for (y_ray, r_ray) to obtain the value of max_dual_ray_infeasibility.
optional double max_dual_ray_infeasibility = 4;
- Returns:
- The maxDualRayInfeasibility.
-
hasDualRayObjective
boolean hasDualRayObjective()The objective of the linear program labeled (1) in the previous paragraph.
optional double dual_ray_objective = 5;
- Returns:
- Whether the dualRayObjective field is set.
-
getDualRayObjective
double getDualRayObjective()The objective of the linear program labeled (1) in the previous paragraph.
optional double dual_ray_objective = 5;
- Returns:
- The dualRayObjective.
-
hasCandidateType
boolean hasCandidateType()Type of the point used to compute the InfeasibilityInformation.
optional .operations_research.pdlp.PointType candidate_type = 6;
- Returns:
- Whether the candidateType field is set.
-
getCandidateType
PointType getCandidateType()Type of the point used to compute the InfeasibilityInformation.
optional .operations_research.pdlp.PointType candidate_type = 6;
- Returns:
- The candidateType.
-