Google OR-Tools v9.11
a fast and portable software suite for combinatorial optimization
|
Static Public Member Functions | |
static final com.google.protobuf.Descriptors.Descriptor | getDescriptor () |
Protected Member Functions | |
com.google.protobuf.GeneratedMessage.FieldAccessorTable | internalGetFieldAccessorTable () |
MPSolverCommonParameters holds advanced usage parameters that apply to any of the solvers we support. All of the fields in this proto can have a value of unspecified. In this case each inner solver will use their own safe defaults. Some values won't be supported by some solvers. The behavior in that case is not defined yet.
Protobuf type operations_research.MPSolverCommonParameters
Definition at line 718 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.MPSolverCommonParameters com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.build | ( | ) |
Definition at line 790 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.MPSolverCommonParameters com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.buildPartial | ( | ) |
Definition at line 799 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.clear | ( | ) |
Definition at line 754 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.clearDualTolerance | ( | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Definition at line 1562 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.clearLpAlgorithm | ( | ) |
Algorithm to solve linear programs. Ask or-core-team@ if you want to know what this does exactly.
optional .operations_research.MPSolverCommonParameters.LPAlgorithmValues lp_algorithm = 4 [default = LP_ALGO_UNSPECIFIED];
Definition at line 1687 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.clearPresolve | ( | ) |
Gurobi and SCIP enable presolve by default. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalBoolean presolve = 5 [default = BOOL_UNSPECIFIED];
Definition at line 1749 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.clearPrimalTolerance | ( | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Definition at line 1372 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.clearRelativeMipGap | ( | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Definition at line 1155 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.clearScaling | ( | ) |
Enable automatic scaling of matrix coefficients and objective. Available for Gurobi and GLOP. Ask or-core-team@ if you want more details.
optional .operations_research.OptionalBoolean scaling = 7 [default = BOOL_UNSPECIFIED];
Definition at line 1815 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.MPSolverCommonParameters com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getDefaultInstanceForType | ( | ) |
Definition at line 785 of file MPSolverCommonParameters.java.
|
static |
Definition at line 723 of file MPSolverCommonParameters.java.
com.google.protobuf.Descriptors.Descriptor com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getDescriptorForType | ( | ) |
Definition at line 780 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDouble com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getDualTolerance | ( | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1469 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDouble.Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getDualToleranceBuilder | ( | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Definition at line 1583 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDoubleOrBuilder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getDualToleranceOrBuilder | ( | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1599 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.MPSolverCommonParameters.LPAlgorithmValues com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getLpAlgorithm | ( | ) |
Algorithm to solve linear programs. Ask or-core-team@ if you want to know what this does exactly.
optional .operations_research.MPSolverCommonParameters.LPAlgorithmValues lp_algorithm = 4 [default = LP_ALGO_UNSPECIFIED];
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1655 of file MPSolverCommonParameters.java.
com.google.ortools.util.OptionalBoolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getPresolve | ( | ) |
Gurobi and SCIP enable presolve by default. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalBoolean presolve = 5 [default = BOOL_UNSPECIFIED];
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1717 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDouble com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getPrimalTolerance | ( | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1283 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDouble.Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getPrimalToleranceBuilder | ( | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Definition at line 1392 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDoubleOrBuilder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getPrimalToleranceOrBuilder | ( | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1407 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDouble com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getRelativeMipGap | ( | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1022 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDouble.Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getRelativeMipGapBuilder | ( | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Definition at line 1186 of file MPSolverCommonParameters.java.
com.google.ortools.linearsolver.OptionalDoubleOrBuilder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getRelativeMipGapOrBuilder | ( | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1212 of file MPSolverCommonParameters.java.
com.google.ortools.util.OptionalBoolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.getScaling | ( | ) |
Enable automatic scaling of matrix coefficients and objective. Available for Gurobi and GLOP. Ask or-core-team@ if you want more details.
optional .operations_research.OptionalBoolean scaling = 7 [default = BOOL_UNSPECIFIED];
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1781 of file MPSolverCommonParameters.java.
boolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.hasDualTolerance | ( | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1454 of file MPSolverCommonParameters.java.
boolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.hasLpAlgorithm | ( | ) |
Algorithm to solve linear programs. Ask or-core-team@ if you want to know what this does exactly.
optional .operations_research.MPSolverCommonParameters.LPAlgorithmValues lp_algorithm = 4 [default = LP_ALGO_UNSPECIFIED];
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1642 of file MPSolverCommonParameters.java.
boolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.hasPresolve | ( | ) |
Gurobi and SCIP enable presolve by default. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalBoolean presolve = 5 [default = BOOL_UNSPECIFIED];
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1704 of file MPSolverCommonParameters.java.
boolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.hasPrimalTolerance | ( | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1269 of file MPSolverCommonParameters.java.
boolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.hasRelativeMipGap | ( | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 997 of file MPSolverCommonParameters.java.
boolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.hasScaling | ( | ) |
Enable automatic scaling of matrix coefficients and objective. Available for Gurobi and GLOP. Ask or-core-team@ if you want more details.
optional .operations_research.OptionalBoolean scaling = 7 [default = BOOL_UNSPECIFIED];
Implements com.google.ortools.linearsolver.MPSolverCommonParametersOrBuilder.
Definition at line 1767 of file MPSolverCommonParameters.java.
|
protected |
Definition at line 729 of file MPSolverCommonParameters.java.
final boolean com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.isInitialized | ( | ) |
Definition at line 878 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.mergeDualTolerance | ( | com.google.ortools.linearsolver.OptionalDouble | value | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Definition at line 1533 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.mergeFrom | ( | com.google.ortools.linearsolver.MPSolverCommonParameters | other | ) |
Definition at line 852 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.mergeFrom | ( | com.google.protobuf.CodedInputStream | input, |
com.google.protobuf.ExtensionRegistryLite | extensionRegistry ) throws java.io.IOException |
Definition at line 883 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.mergeFrom | ( | com.google.protobuf.Message | other | ) |
Definition at line 843 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.mergePrimalTolerance | ( | com.google.ortools.linearsolver.OptionalDouble | value | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Definition at line 1344 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.mergeRelativeMipGap | ( | com.google.ortools.linearsolver.OptionalDouble | value | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Definition at line 1116 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setDualTolerance | ( | com.google.ortools.linearsolver.OptionalDouble | value | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Definition at line 1487 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setDualTolerance | ( | com.google.ortools.linearsolver.OptionalDouble.Builder | builderForValue | ) |
Tolerance for dual feasibility. For SCIP and Gurobi this is the feasibility tolerance for reduced costs in LP solution: reduced costs must all be smaller than this value in the improving direction in order for a model to be declared optimal. Not supported for other solvers.
optional .operations_research.OptionalDouble dual_tolerance = 3;
Definition at line 1511 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setLpAlgorithm | ( | com.google.ortools.linearsolver.MPSolverCommonParameters.LPAlgorithmValues | value | ) |
Algorithm to solve linear programs. Ask or-core-team@ if you want to know what this does exactly.
optional .operations_research.MPSolverCommonParameters.LPAlgorithmValues lp_algorithm = 4 [default = LP_ALGO_UNSPECIFIED];
value | The lpAlgorithm to set. |
Definition at line 1669 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setPresolve | ( | com.google.ortools.util.OptionalBoolean | value | ) |
Gurobi and SCIP enable presolve by default. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalBoolean presolve = 5 [default = BOOL_UNSPECIFIED];
value | The presolve to set. |
Definition at line 1731 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setPrimalTolerance | ( | com.google.ortools.linearsolver.OptionalDouble | value | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Definition at line 1300 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setPrimalTolerance | ( | com.google.ortools.linearsolver.OptionalDouble.Builder | builderForValue | ) |
Tolerance for primal feasibility of basic solutions: this is the maximum allowed error in constraint satisfiability. For SCIP this includes integrality constraints. For Gurobi it does not, you need to set the custom parameter IntFeasTol.
optional .operations_research.OptionalDouble primal_tolerance = 2;
Definition at line 1323 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setRelativeMipGap | ( | com.google.ortools.linearsolver.OptionalDouble | value | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Definition at line 1050 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setRelativeMipGap | ( | com.google.ortools.linearsolver.OptionalDouble.Builder | builderForValue | ) |
The solver stops if the relative MIP gap reaches this value or below. The relative MIP gap is an upper bound of the relative distance to the optimum, and it is defined as: abs(best_bound - incumbent) / abs(incumbent) [Gurobi] abs(best_bound - incumbent) / min(abs(best_bound), abs(incumbent)) [SCIP] where "incumbent" is the objective value of the best solution found so far (i.e., lowest when minimizing, highest when maximizing), and "best_bound" is the tightest bound of the objective determined so far (i.e., highest when minimizing, and lowest when maximizing). The MIP Gap is sensitive to objective offset. If the denominator is 0 the MIP Gap is INFINITY for SCIP and Gurobi. Of note, "incumbent" and "best bound" are called "primal bound" and "dual bound" in SCIP, respectively. Ask or-core-team@ for other solvers.
optional .operations_research.OptionalDouble relative_mip_gap = 1;
Definition at line 1084 of file MPSolverCommonParameters.java.
Builder com.google.ortools.linearsolver.MPSolverCommonParameters.Builder.setScaling | ( | com.google.ortools.util.OptionalBoolean | value | ) |
Enable automatic scaling of matrix coefficients and objective. Available for Gurobi and GLOP. Ask or-core-team@ if you want more details.
optional .operations_research.OptionalBoolean scaling = 7 [default = BOOL_UNSPECIFIED];
value | The scaling to set. |
Definition at line 1796 of file MPSolverCommonParameters.java.