optimality

Optimality refers to the quality of being the best according to a specified criterion within a defined set of feasible choices. In mathematical optimization, an optimal solution is a point that maximizes or minimizes an objective function subject to constraints. The criterion and the feasible set together determine what counts as optimal.

Local versus global: a local optimum is best in a neighborhood but not necessarily overall. A global optimum is the best possible value over the entire feasible region. Pareto optimality, used in economics and welfare theory, describes allocations where no one can be made better off without making someone else worse off.

Existence and conditions: if the objective is continuous and the feasible set is compact, a minimum and maximum exist (Weierstrass). Necessary conditions include first-order (gradient zero) and second-order conditions. For constrained problems, Lagrange multipliers provide optimality conditions; Karush-Kuhn-Tucker conditions extend to inequality constraints. If the problem is convex, any local optimum is global. For nonconvex problems, multiple local optima may exist.

Methods and computation: gradient-based methods seek stationary points; convex optimization yields efficient algorithms with guarantees. Linear programming and integer programming are special cases. Dynamic programming uses optimal substructure and Bellman principles. In stochastic settings, optimality can refer to Bayes optimal estimators or risk minimization under uncertainty.

Applications and interpretation: optimality concepts appear in economics, engineering, computer science, operations research, and statistics. They depend on the chosen objective and constraints, and may involve trade-offs or approximations. Optima may be unique or form a set, and practical models may yield near-optimal or robust solutions.