An auxiliary variable introduced to convert an optimization problem with smooth equality constraints into a stationarity condition for an augmented Lagrangian function, enabling simultaneous solution for variables and constraints.

Principle

Form the Lagrangian L(x,λ)=f(x)+λᵀ g(x) for objective f and equality constraints g(x)=0; stationary points of L with respect to x and λ yield candidate constrained extrema under regularity conditions.

Demonstration

Maximize f(x,y)=xy subject to x+y=1. L(x,y,λ)=xy+λ(x+y−1). Solve ∂L/∂x= y+λ=0, ∂L/∂y= x+λ=0, ∂L/∂λ= x+y−1=0 to obtain critical points and λ that enforce the constraint.

Misapplication

Applying the method when constraint gradients are zero at the solution (violating regularity) or failing to check second‑order conditions and boundary behavior can yield spurious solutions.

Consequence

Gives necessary conditions to find constrained extrema and interprets multipliers as sensitivities of the optimal value to infinitesimal changes in constraint right‑hand sides (shadow prices).

Reversal

One can eliminate multipliers by explicitly solving constraints (substitution) when convenient; multipliers reappear as Lagrange undetermined constants in implicit elimination.

Boundary

Requires differentiable objective and constraints and applies to equality constraints; inequality constraints require KKT conditions and additional complementary slackness considerations.

Semantic Tension

Contrasted with penalty or barrier methods: multipliers enforce constraints exactly at stationarity and yield dual information, while penalty methods approximate constraints and may distort conditioning.

Synthesis

A Lagrange multiplier is an auxiliary variable that enforces smooth equality constraints by forming a Lagrangian whose stationary conditions provide candidate constrained extrema and sensitivity information.