Lagrange Multipliers- Two-Constraint Problems
What Are Lagrange Multipliers with Two Constraints?
Lagrange multipliers help you find the maximum or minimum of a function when something is holding you back. Usually, you deal with one constraint. But real problems often have two restrictions at once.
That's where two-constraint Lagrange multipliers come in. You optimize f(x,y,z) subject to g(x,y,z) = 0 and h(x,y,z) = 0 simultaneously.
It's not twice as hard as one constraint. The method just scales up in a predictable way.
The Setup
You want to extremize:
f(x, y, z) — your objective function
Subject to two constraints:
g(x, y, z) = 0 and h(x, y, z) = 0
The Lagrangian becomes:
L = f + λg + μh
Two multipliers now. λ (lambda) for the first constraint, μ (mu) for the second.
The System of Equations
Take partial derivatives and set them to zero:
∂L/∂x = 0 → ∂f/∂x + λ∂g/∂x + μ∂h/∂x = 0
∂L/∂y = 0 → ∂f/∂y + λ∂g/∂y + μ∂h/∂y = 0
∂L/∂z = 0 → ∂f/∂z + λ∂g/∂z + μ∂h/∂z = 0
Plus your two constraints:
g(x, y, z) = 0
h(x, y, z) = 0
Five equations. Five unknowns. (x, y, z, λ, μ)
Solve the system. That's it.
How To Solve Two-Constraint Problems
Step 1: Identify Your Functions
Write down f, g, and h clearly. Messing this up kills everything downstream.
Step 2: Build the Lagrangian
L = f(x,y,z) + λg(x,y,z) + μh(x,y,z)
Don't forget the multipliers.
Step 3: Compute All Partial Derivatives
You need ∂L/∂x, ∂L/∂y, ∂L/∂z. Do each one carefully.
Step 4: Set Up the Full System
Write all five equations. Check twice.
Step 5: Solve
You have five unknowns. Use substitution, elimination, or linear algebra. Whatever gets you to the answer.
Step 6: Evaluate
Plug candidate points into f to find which gives max/min. If you need definiteness, use the bordered Hessian.
Example Problem
Find the maximum of f(x,y,z) = x + y + z on the intersection of:
x² + y² + z² = 1 (unit sphere)
x + y = 0 (plane)
Here: g = x² + y² + z² - 1, h = x + y
The Lagrangian:
L = x + y + z + λ(x² + y² + z² - 1) + μ(x + y)
Partial derivatives:
∂L/∂x = 1 + 2λx + μ = 0
∂L/∂y = 1 + 2λy + μ = 0
∂L/∂z = 1 + 2λz = 0
Constraints:
x² + y² + z² = 1
x + y = 0
From ∂L/∂x and ∂L/∂y: subtract them → 2λ(x - y) = 0
Either λ = 0 or x = y
If λ = 0: then 1 + μ = 0 and 1 + μ = 0, so μ = -1. And 1 + 2λz = 1 = 0. Contradiction. Toss this branch.
So x = y. Combined with x + y = 0 → x = y = 0.
Then z² = 1 from the sphere, so z = ±1.
Check: (0, 0, 1) gives f = 1. (0, 0, -1) gives f = -1.
Maximum is 1 at (0, 0, 1).
Quick Comparison Table
| Aspect | One Constraint | Two Constraints |
|---|---|---|
| Variables | 2 or 3 typically | 3 or more |
| Multipliers | λ only | λ and μ |
| Equations | 3 (2 partial + 1 constraint) | 5 (3 partial + 2 constraints) |
| Geometry | Curve on surface | Point(s) on curve intersection |
Common Mistakes
- Forgetting a constraint — you'll get the wrong answer every time
- Algebra errors in partial derivatives — double-check signs
- Not testing all candidate points — multiple solutions appear often
- Confusing which constraint gets which multiplier — label them clearly
When Two Constraints Actually Show Up
Physics problems with conserved quantities. Economics with two resource limits. Engineering with multiple design constraints. The method doesn't change — you just have more equations to solve.
If your constraints are dependent (one is a multiple of the other), you effectively have one constraint. Watch for this.
The Bottom Line
Two-constraint Lagrange multipliers are straightforward once you see the pattern. Build the Lagrangian with two multipliers, take three partial derivatives, add two constraints, solve five equations.
The algebra gets messier. The logic doesn't change at all from the one-constraint case.