<< Chapter < Page | Chapter >> Page > |
In constrained optimization, we look to minimize or maximize an objective function only over a set of inputs that can be written in the following form:
In words, is a set described by set of equalities and inequalities on . The constraints are said to be equality constraints , and the constraints are said to be inequality constraints .
Example 1 Minimize subject to , and , drawn on [link] .
In this optimization problem, the feasible set can be written in terms of one equality constraint, , and four inequality constraints:
Definition 1 The points are called feasible points , and the set is called the feasible set .
In the sequel, we will assume that , , are continuous and Fréchet-differentiable (continuous gradients).
We will first consider problems where the feasible set can be expressed in terms of equality constraints only.
For a given feasible point , the tangent space gives us the set of directions in which one can move from while still staying within the feasible set . The two examples below show tangent spaces in the cases where the set correspond to a curve in and a nonlinear manifold in .
The tangent space at can be expressed formally in terms of the derivatives of the equality constraints.
Definition 2 The tangent space to the feasible set with equality constraints at a feasible point is given by
Requiring all the directional derivatives of the equality constraints to be zero in the direction of the tangent guarantees that the value of the equality constraints remains at zero, therefore guaranteeing that remains a feasible point, i.e., .
Definition 3 A point satisfying the constraints is said to be a regular point if the linear functionals (i.e., the derivatives of the equality constraints) are linearly independent on .
Theorem 1 If is an extremum of subject to constraints and is a regular point of then for any such that for all , we must have .
One can rewrite this theorem in terms of gradients as: if , then . More intuition can be obtained by defining the translated tangent space at as follows:
Thus, the theorem above can be written as: if , then . In words, we expect for the derivative of the objective function at the constrained extremum to be zero-valued in the directions in which we can move from and remain in the feasible set — that is, in the directions . Thus, we can write that the constrained optimum must obey , which implies .
Notification Switch
Would you like to follow the 'Signal theory' conversation and receive update notifications?