Feasible Set

A feasible set is the subset of a decision space that satisfies all constraints imposed on an optimization problem. In geometric terms, it is the intersection of all constraint surfaces—the region where the problem's rules permit solutions to exist. The shape, convexity, and topology of the feasible set determine whether an optimum can be found efficiently, whether it is unique, and whether local optimality implies global optimality.

The feasible set is often treated as a passive container within which optimization occurs. This is a mistake. The feasible set is an active participant in the optimization process: its curvature determines convergence rates, its vertices define the combinatorial structure of discrete optimization, and its dimensionality relative to the ambient space measures the problem's effective complexity. In convex optimization, the feasible set's convexity is not a convenience but a guarantee—without it, the problem belongs to a different computational complexity class.

The feasible set is where the rubber of constraint meets the road of possibility. Every optimization algorithm is, in the end, a negotiation with the geometry of the feasible set—and the set, like any good negotiator, dictates the terms. \n\n== See Also ==\n\n* Active Constraint — a constraint that is satisfied with equality at the optimum and therefore actively shapes the local geometry of the feasible set.\n* Slack Variable — an auxiliary variable introduced to convert an inequality constraint into an equality by measuring the distance from the constraint boundary.