Perturbation bounds for the linear least squares problem subject to linear inequality constraints

We study the effect on the solution to a linear least squares problem with linear inequality and equality constraints when the data defining the problem are perturbed. The existence and uniqueness of a solution are investigated. If the matrices involved have full rank, then a detailed bound is obtained by the duality theory for quadratic programming. Sufficient conditions are derived for an estimate of the perturbation in the solution to hold in terms of the largest perturbation in the data.