getbound_per_relax: getbound_per_relax

In seanmtelliott/RCBRpartial: Partial Identification of Random Coefficient Binary Response Model

Usage

getbound_per_relax(obj, Acons, bcons, Aextra, bextra, pervar, verb = 1)

Arguments

obj

defines the linear coefficients of the linear objective function

Acons

defines the linear constraints in the form of Acons \item bconsdefines the linear constraints in the form of Acons \item Aextradefines the relaxed linear equality constraints in the form Aextra \item bextradefines the amount of relaxation of the equality constraints. \item pervardefines the perturbation as in Cho and Russell (2020). optimal value of the objective functions for the lower_minus, lower_plus, upper_minus, upper_plus LP problems. Definition see Cho and Russell (2020). phat are the associated optimal solutions and status is the solution status of the optimization problem. #' @examples In the original LP, we allow relaxation of the IV assumption: Aextra Aextra Aextra while maintaining the equality constraints Acons And then we solve the four perturbed LP problems as in Cho and Russell (2020) using Rmosek. Original LP is defined as ## min/max_x obj'x s.t. Acons %*% x - bcons = 0 Aextra %*% x <= bextra Aextra %*% x >= -bextra We then define the four perturbed LP as in Cho and Russell (2020).

seanmtelliott/RCBRpartial documentation built on Dec. 22, 2021, 11:16 p.m.