 Named after Joseph Louis Lagrange.  Lagrange multipliers provides a strategy for finding

the maxima and minima of a function subject to constraints.  This method gives a set of necessary conditions to identify optimal points of equality constrained optimization problems.  Done by converting a constrained problem to an equivalent unconstrained problem with the help of certain unspecified parameters known as Lagrange multipliers.

 The method of Lagrange multipliers is a powerful tool for solving this class of problems.  Difficulties often arise when one wishes to maximize or minimize a function subject to fixed outside conditions or constraints. One of the most common problems in calculus is that of finding maxima or minima of a function. .

xn) then you will get n equations with n + 1 unknowns (i.. x 1. Original problem is rewritten as: minimize L(x. . (If there are n variables (i.. is called the Lagrange Multiplier.e.λ h1(x) ...e. . The constant. ..  Calculate x by using the just found value for λ. n variables xi and one Lagrangian multiplier λ)  Express all xi in terms of Langrangian multiplier λ  Plug x in terms of λ in constraint h1(x) = 0 and solve λ. λ) with respect to xi and set them equal to zero. λ) = f(x) .  Take derivatives of L(x.


y)/δx = 2 x δg(x. y)/δy = -3 δg(x. y) = x² + y²-136 Thus. f(x. δf(x. y) = 5x – 3y constraint g(x. y)/δx = 5 δ f(x. Here. y)/δy = 2 y .

x=-10 and y= 6 To determine if we have maximums or minimums we just need to plug these into the function.• Now. y)/δy = λ δg(x.-6) . y)/δx δf(x. y)/δx = λ δg(x. applying δf(x. 5 = 2 λx -3 = 2 λy x² + y²=136 (constraint) Solving these equations x = 5/2 λ y = -3/2 λ Plugging these into the constraint we get λ=±¼ If λ = ¼ then. x=10 and y=-6 if λ=-1/4 then.6) = -68 Minimum at(-10.6) f(10. Here are the minimum and maximum values of the function..-6) = 68 Maximum at(10. y)/δy We get following set of equations . f(-10.

z) = x y + y z + z x-32 Thus. z)/δx = (y + z) δg(x. y. z)/δz = y x δg(x. we get. δf(x.e. y.. y. y. y. y. x y + y z + z x = 32 g(x . y. z)/δy = (x + z) δg(x.y . f(x. z)/δx = y z δf(x. Here. z) = xyz constraint 2(x y + y z + z x) = 64 i. z)/δy = x z δf(x. z)/δz = (y + x) (volume of the box) (total surface area) .

y... z)/δy = λ δg(x. z)/δz We get following set of equations . y. y. y. z)/δz = λ δg(x. y.266 .266 We can say that we will get a maximum volume if the dimensions are x = y = z = 3. x = y = z = 3. y z = λ (y + z) x z = λ (x + z) y x = λ (y + x) x y + y z + z x = 32 (constraint) Applying either elimination or substitution method we now solve the set of equations thus obtained. y. z)/δx δf(x. z)/δy δf(x. Now. applying δf(x. Thus. z)/δx = λ δg(x.

z ≥ 0.  Find the maximum and minimum of f(x. Assume that x. y. y. Find the maximum and minimum values of f(x.  Find the maximum and minimum values of f(x. y. Consider the following……. y) = 4x² + 10y² on the disk while x² + y²≤ 4. . z) = x y z subject to the constraint x + y + z = 1. z) = 4y – 2z subject to constraints 2x – y – z = 2 and x² + y² = 1.

from the first step into f(x. Solve the following system of equations. f(P) = λ1 g1(P) + λ2 g2(P) + … + λn g n(P) g1(P) = k1 g2(P) = k2 . provided they exist. λ2. g3=0.. . λn is called the Lagrange Multiplier. g n(P) = k n • Plug in all solutions. • The constants. ………. g n=0 are n number of constraints then. …. If g1=0. g2=0. . (x. λ1 . . y.. z).. z) and identify the minimum and maximum values. y.

x3.x3.x4.x5)/δx4 =2x4 δf(x1.x5)/δx3 =2x3 δf(x1.x3.x5)/δx5 =2x5 .x3.x3.x5)=x3-2x4+x5-6 Thus.x3.x3.x2.x5)/δx2 =2x2 δf(x1.x2.x4. δf(x1.x4.x4.x4.x5)/δx1 =2x1 δf(x1.x5)= x12+x22+x32+x42+x52 Constraints g1(x1.x2.x2.x2.x2. f(x1.x4.x3.Here.x2.x2.x4.x5)=x1+2x2+x3-1 g2(x1.x4.

x4.x3.x2.x4.x2.x5)/δx3 =1 δg1 (x1.x3.x2.x5)/δx4 =-2 δg2 (x1.x2.x5)/δx1 = λ δg1(x1.x4.x3.x4.x2.x5)/δx3 =1 δg2 (x1.x5)/δ x1 .x5)/δx4 =0 δg1 (x1.x2.x2.x4.x5)/δ x1 + µ δg2(x1.x3.x4.x3.x4.x2.x4.x2.x3.x4.x4.x5)/δx2 =2 δg1 (x1. δf(x1.x4.x3.x5)/δ x1 =0 δg2 (x1.x4.x2.x2.x4.x3.x3.x2.x5)/δx5 =1 On applying.x5)/δx2 =0 δg2 (x1.x2.δg1(x1.x3.x5)/δ x1 =1 δg1 (x1.x3.x5)/δx5 =0 δg2(x1.x3.x3.

x3.x4.x5)/δx4= λ δg1(x1. 2x4 -2 µ =0 .x2.x3.x2.x4.x5)/δ x3 δf(x1.x3.x5)/δ x3 + µ δg2(x1.x4.x5)/δ x5 We get following set of equations . 2x1 + λ=0 .x4.x3.x2.x5)/δx2 + µ (x1. 2x5 + µ =0 x1+2x2+x3=1 .x2.x5)/δ x4 δf(x1.x4.x5)/δx2= λ δg1 (x1.x4.x2.x3. 2x3 + λ + µ =0 .x3.x2.δf(x1.x5)/δx3= λ δg1(x1.x3..x4.x3. 2x2 +2 λ=0 .x3.x4.x4.x3.x4.x3. x3-2x4+x5=6 (constraint) δg2 .x4.x4.x2.x5)/δ x5 + µ δg2(x1.x2.x2.x5)/δ x4 + µ δg2(x1.x2.x2.x2.x3.x5)/δx2 δf(x1.x5)/δx5= λ δg1(x1.

1. Here .1)=6. x3=x5=1 and x4=-2 To determine minimums we just need to plug these into the function.-2. λ=0 Plugging these into the equations we get x1=x2=0. f(0.On Solving these equations µ=-2. .the minimum values of the function.0.

z)/δx =0 δf(x .y . δf(x . f(x .y .y .y . y .z)/δy =4 δf(x .z)/δy =-1 .Here. z)=4y-2z Constraint.y .z)= x² + y² -1 Thus.z)/δz =-2 δg1(x .z)/δx =2 δg1(x . g1(x .z)=2x-y-z-2 g2(x .y .y .

2x – y – z = 2.y .y .λ=-2. .y .z)/δx δf(x .y . 2 λ+2x µ=0.y . (constraint) x² + y² = 1 (constraint) .z)/δx= λ δg1 (x .y .z)/δy= λ δg1 (x .y .λ+2y µ=4.z)/δz= λ δg1 (x .y .δg1(x .z)/δz =-1 δg2(x .y .y .z)/δz + µ δg2 (x .z)/δy+ µ δg2 (x .z)/δz=0 On applying.z)/δy δf(x .z)/δx=2x δg2(x . δf(x .y . .z)/δx + µ δg2 (x ..z)/δz We get following set of equations .z)/δy =2y δg2(x .y .y .

y=0.y=-0. f(0. µ=+5.-5 Plugging these into the equations we get If µ=+5 .2)=10.z=-4.2 To determine if we have maximums or minimums we just need to plug these into the function.then x=0.2) f(-0.6.On Solving these equations… λ=2.z= minimum at(0.- maximum at(-0.2) .2 and If µ=-5 . x=-0.8.6.-0. Here are the minimum and maximum values of the function.-4.6.-0.0.

 Kuhn and Tucker extended the Lagrangian theory to include the general classical single-objective nonlinear programming problem: . Lagrangian multipliers are very useful in sensitivity analyses.  Lagrangian multipliers require equalities.  Setting the derivatives of L to zero may result in finding a saddle point. So a conversion of inequalities is necessary. Additional checks are always useful.