Pattern Recognition, 4th Edition

C.4.2. Inequality Constraints

The general problem can be cast as follows:(C.29)Each one of the constraints defines a region in R¹. The intersection of all these regions defines the area in which the constrained minimum, θ^*, must lie. This is known as the feasible region and the points in it (candidate solutions) as feasible points. The type of the constraints control the type of the feasible region, that is, whether it is convex or concave. At this point, it will not harm us to recall a few definitions.

Convex functions. A function f (θ)is called ...