Reference

MathOptNLPModel(model, hessian=true, name="Generic")

Construct a MathOptNLPModel from a JuMP model.

hessian should be set to false for multivariate user-defined functions registered without hessian.

MathOptNLSModel(model, F, hessian=true, name="Generic")

Construct a MathOptNLSModel from a JuMP model and a container of JuMP GenericAffExpr (generated by @expression) and NonlinearExpression (generated by @NLexpression).

hessian should be set to false for multivariate user-defined functions registered without hessian.

NonLinearStructure

Structure containing Jacobian and Hessian structures of nonlinear constraints:

• nnln: number of nonlinear constraints
• nl_lcon: lower bounds of nonlinear constraints
• nl_ucon: upper bounds of nonlinear constraints
• jac_rows: row indices of the Jacobian in Coordinate format (COO) format
• jac_cols: column indices of the Jacobian in COO format
• nnzj: number of non-zero entries in the Jacobian
• hess_rows: row indices of the Hessian in COO format
• hess_cols: column indices of the Hessian in COO format
• nnzh: number of non-zero entries in the Hessian

Oracles

Structure containing nonlinear oracles data:

• oracles: vector of tuples (MOI.VectorOfVariables, _VectorNonlinearOracleCache)
• ncon: number of scalar constraints represented by all oracles
• lcon: lower bounds of oracle constraints
• ucon: upper bounds of oracle constraints
• nnzj: number of non-zero entries in the Jacobian of all oracles
• nnzh: number of non-zero entries in the Hessian of all oracles
• nzJ: buffer to store the nonzeros of the Jacobian for all oracles (needed for the functions jprod and jtprod)
• nzH: buffer to store the nonzeros of the Hessian for all oracles (needed for the function hprod)
• hessianoraclessupported: support of the Hessian for all oracles

add_constraint_model(Fmodel, Fi)

Add the nonlinear constraint Fi == 0 to the model Fmodel. If Fi is an Array, then we iterate over each component.

coo_sym_add_mul!(rows, cols, vals, x, y, α)

Perform the update y ← y + α * A * x where A is a symmetric matrix in COO format given by (rows, cols, vals). Only one triangle of A should be passed.

coo_sym_dot(rows, cols, vals, x, y)

Compute the product xᵀAy of a symmetric matrix A given by (rows, cols, vals). Only one triangle of A should be passed.

coo_unsym_add_mul!(transpose, rows, cols, vals, x, y, α)

Performs the update y ← y + α * op(A) * x, where A is an unsymmetric matrix in COO format given by (rows, cols, vals). If transpose == true, then op(A) = Aᵀ; otherwise, op(A) = A.

parser_MOI(moimodel, index_map, nvar)

Parse linear constraints of a MOI.ModelLike.

parser_NL(nlp_data; hessian)

Parse nonlinear constraints of an nlp_data.

Returns:

• nlcon: NonLinearStructure containing Jacobian and Hessian structures

parser_SAF(fun, set, linrows, lincols, linvals, nlin, lin_lcon, lin_ucon, index_map)

Parse a ScalarAffineFunction fun with its associated set. linrows, lincols, linvals, lin_lcon and lin_ucon are updated.

parser_SQF(fun, set, nvar, qcons, quad_lcon, quad_ucon, index_map)

Parse a ScalarQuadraticFunction fun with its associated set. qcons, quad_lcon, quad_ucon are updated.

parser_VAF(fun, set, linrows, lincols, linvals, nlin, lin_lcon, lin_ucon, index_map)

Parse a VectorAffineFunction fun with its associated set. linrows, lincols, linvals, lin_lcon and lin_ucon are updated.

parser_VQF(fun, set, nvar, qcons, quad_lcon, quad_ucon, index_map)

Parse a VectorQuadraticFunction fun with its associated set. qcons, quad_lcon, quad_ucon are updated.

parser_linear_expression(cmodel, nvar, index_map, F)

Parse linear expressions of type VariableRef and GenericAffExpr{Float64,VariableRef}.

parser_nonlinear_expression(cmodel, nvar, F; hessian)

Parse nonlinear expressions of type NonlinearExpression.

parser_objective_MOI(moimodel, nvar, index_map)

Parse linear and quadratic objective of a MOI.ModelLike.

oracles = parser_oracles(moimodel)

Parse nonlinear oracles of a MOI.ModelLike.

parser_variables(model)

Parse variables informations of a MOI.ModelLike.

replace!(ex, x)

Walk the expression ex and substitute in the actual variables x.