Tools

Functions evaluations

After calling one the API functions to get a function value, the number of times that function was called is stored inside the NLPModel. For instance

using ADNLPModels, LinearAlgebra, NLPModels
nlp = ADNLPModel(x -> dot(x, x), zeros(2))
for i = 1:100
   obj(nlp, rand(2))
end
neval_obj(nlp)

Some counters are available for all models, some are specific. In particular, there are additional specific counters for the nonlinear least squares models (the ones with residual below).

Counter	Description
`neval_obj`	Objective
`neval_grad`	Gradient
`neval_cons`	Constraints
`neval_cons_lin`	Linear constraints
`neval_cons_nln`	Nonlinear constraints
`neval_jcon`	One constraint - unused
`neval_jgrad`	Gradient of one constraints - unused
`neval_jac`	Jacobian
`neval_jac_lin`	Linear constraints Jacobian
`neval_jac_nln`	Nonlinear constraints Jacobian
`neval_jprod`	Product of Jacobian and vector
`neval_jprod_lin`	Product of linear constraints Jacobian and vector
`neval_jprod_nln`	Product of nonlinear constraints Jacobian and vector
`neval_jtprod`	Product of transposed Jacobian and vector
`neval_jtprod_lin`	Product of transposed linear constraints Jacobian and vector
`neval_jtprod_nln`	Product of transposed nonlinear constraints Jacobian and vector
`neval_hess`	Hessian
`neval_hprod`	Product of Hessian and vector
`neval_jhess`	Individual Lagrangian Hessian evaluations
`neval_jhprod`	Product of Hessian of j-th function and vector
`neval_residual`	Residual function of nonlinear least squares model
`neval_jac_residual`	Jacobian of the residual
`neval_jprod_residual`	Product of Jacobian of residual and vector
`neval_jtprod_residual`	Product of transposed Jacobian of residual and vector
`neval_hess_residual`	Sum of Hessians of residuals
`neval_jhess_residual`	Hessian of a residual component
`neval_hprod_residual`	Product of Hessian of a residual component and vector

To get the sum of all counters except cons, jac, jprod and jtprod called for a problem, use sum_counters.

using ADNLPModels, LinearAlgebra, NLPModels
nlp = ADNLPModel(x -> dot(x, x), zeros(2))
obj(nlp, rand(2))
grad(nlp, rand(2))
sum_counters(nlp)

Querying problem type

There are some utility functions for querying the problem type:

has_bounds: True when not all variables are free.
bound_constrained: True for problems with bounded variables and no other constraints.
equality_constrained: True when problem is constrained only by equalities.
has_equalities: True when problem has at least one equality constraint.
inequality_constrained: True when problem is constrained by inequalities.
has_inequalities: True when problem has at least one inequality constraint that isn't a bound.
linearly_constrained: True when problem is constrained by equalities or inequalities known to be linear.
unconstrained: True when problem is not constrained.