Lagrange multiplier

Last updated

In mathematical optimization, the method of Lagrange multipliers is a strategy for finding the local maxima and minima of a function subject to equation constraints (i.e., subject to the condition that one or more equations have to be satisfied exactly by the chosen values of the variables). [1] It is named after the mathematician Joseph-Louis Lagrange.

Contents

Summary and rationale

The basic idea is to convert a constrained problem into a form such that the derivative test of an unconstrained problem can still be applied. The relationship between the gradient of the function and gradients of the constraints rather naturally leads to a reformulation of the original problem, known as the Lagrangian function. [2] In the general case, the Lagrangian is defined as

for functions ; is called the Lagrange multiplier.

In simple cases, where the dot product is defined as the inner product, the Lagrangian is

The method can be summarized as follows: in order to find the maximum or minimum of a function subjected to the equality constraint , find the stationary points of considered as a function of and the Lagrange multiplier . This means that all partial derivatives should be zero, including the partial derivative with respect to . [3]

and

or equivalently

and

The solution corresponding to the original constrained optimization is always a saddle point of the Lagrangian function, [4] [5] which can be identified among the stationary points from the definiteness of the bordered Hessian matrix. [6]

The great advantage of this method is that it allows the optimization to be solved without explicit parameterization in terms of the constraints. As a result, the method of Lagrange multipliers is widely used to solve challenging constrained optimization problems. Further, the method of Lagrange multipliers is generalized by the Karush–Kuhn–Tucker conditions, which can also take into account inequality constraints of the form for a given constant .

Statement

The following is known as the Lagrange multiplier theorem. [7]

Let be the objective function, be the constraints function, both belonging to (that is, having continuous first derivatives). Let be an optimal solution to the following optimization problem such that, for the matrix of partial derivatives , :

Then there exists a unique Lagrange multiplier such that (Note that this is a somewhat conventional thing where is clearly treated as a column vector to ensure that the dimensions match. But, we might as well make it just a row vector without taking the transpose).

The Lagrange multiplier theorem states that at any local maximum (or minimum) of the function evaluated under the equality constraints, if constraint qualification applies (explained below), then the gradient of the function (at that point) can be expressed as a linear combination of the gradients of the constraints (at that point), with the Lagrange multipliers acting as coefficients. [8] This is equivalent to saying that any direction perpendicular to all gradients of the constraints is also perpendicular to the gradient of the function. Or still, saying that the directional derivative of the function is 0 in every feasible direction.

Single constraint

Figure 1: The red curve shows the constraint g(x, y) = c. The blue curves are contours of f(x, y). The point where the red constraint tangentially touches a blue contour is the maximum of f(x, y) along the constraint, since d1 > d2 . LagrangeMultipliers2D.svg
Figure 1: The red curve shows the constraint g(x, y) = c. The blue curves are contours of f(x, y). The point where the red constraint tangentially touches a blue contour is the maximum of f(x, y) along the constraint, since d1 > d2.

For the case of only one constraint and only two choice variables (as exemplified in Figure 1), consider the optimization problem

(Sometimes an additive constant is shown separately rather than being included in , in which case the constraint is written as in Figure 1.) We assume that both and have continuous first partial derivatives. We introduce a new variable () called a Lagrange multiplier (or Lagrange undetermined multiplier) and study the Lagrange function (or Lagrangian or Lagrangian expression) defined by

where the term may be either added or subtracted. If is a maximum of for the original constrained problem and then there exists such that () is a stationary point for the Lagrange function (stationary points are those points where the first partial derivatives of are zero). The assumption is called constraint qualification. However, not all stationary points yield a solution of the original problem, as the method of Lagrange multipliers yields only a necessary condition for optimality in constrained problems. [9] [10] [11] [12] [13] Sufficient conditions for a minimum or maximum also exist, but if a particular candidate solution satisfies the sufficient conditions, it is only guaranteed that that solution is the best one locally – that is, it is better than any permissible nearby points. The global optimum can be found by comparing the values of the original objective function at the points satisfying the necessary and locally sufficient conditions.

The method of Lagrange multipliers relies on the intuition that at a maximum, f(x, y) cannot be increasing in the direction of any such neighboring point that also has g = 0. If it were, we could walk along g = 0 to get higher, meaning that the starting point wasn't actually the maximum. Viewed in this way, it is an exact analogue to testing if the derivative of an unconstrained function is 0, that is, we are verifying that the directional derivative is 0 in any relevant (viable) direction.

We can visualize contours of f given by f(x, y) = d for various values of d, and the contour of g given by g(x, y) = c.

Suppose we walk along the contour line with g = c . We are interested in finding points where f almost does not change as we walk, since these points might be maxima.

There are two ways this could happen:

  1. We could touch a contour line of f, since by definition f does not change as we walk along its contour lines. This would mean that the tangents to the contour lines of f and g are parallel here.
  2. We have reached a "level" part of f, meaning that f does not change in any direction.

To check the first possibility (we touch a contour line of f), notice that since the gradient of a function is perpendicular to the contour lines, the tangents to the contour lines of f and g are parallel if and only if the gradients of f and g are parallel. Thus we want points (x, y) where g(x, y) = c and

for some

where

are the respective gradients. The constant is required because although the two gradient vectors are parallel, the magnitudes of the gradient vectors are generally not equal. This constant is called the Lagrange multiplier. (In some conventions is preceded by a minus sign).

Notice that this method also solves the second possibility, that f is level: if f is level, then its gradient is zero, and setting is a solution regardless of .

To incorporate these conditions into one equation, we introduce an auxiliary function

and solve

Note that this amounts to solving three equations in three unknowns. This is the method of Lagrange multipliers.

Note that implies as the partial derivative of with respect to is

To summarize

The method generalizes readily to functions on variables

which amounts to solving n + 1 equations in n + 1 unknowns.

The constrained extrema of f are critical points of the Lagrangian , but they are not necessarily local extrema of (see Example 2 below).

One may reformulate the Lagrangian as a Hamiltonian, in which case the solutions are local minima for the Hamiltonian. This is done in optimal control theory, in the form of Pontryagin's minimum principle.

The fact that solutions of the method of Lagrange multipliers are not necessarily extrema of the Lagrangian, also poses difficulties for numerical optimization. This can be addressed by minimizing the magnitude of the gradient of the Lagrangian, as these minima are the same as the zeros of the magnitude, as illustrated in Example 5: Numerical optimization.

Multiple constraints

Figure 2: A paraboloid constrained along two intersecting lines. As wiki lgm parab.svg
Figure 2: A paraboloid constrained along two intersecting lines.
Figure 3: Contour map of Figure 2. As wiki lgm levelsets.svg
Figure 3: Contour map of Figure 2.

The method of Lagrange multipliers can be extended to solve problems with multiple constraints using a similar argument. Consider a paraboloid subject to two line constraints that intersect at a single point. As the only feasible solution, this point is obviously a constrained extremum. However, the level set of is clearly not parallel to either constraint at the intersection point (see Figure 3); instead, it is a linear combination of the two constraints' gradients. In the case of multiple constraints, that will be what we seek in general: The method of Lagrange seeks points not at which the gradient of is multiple of any single constraint's gradient necessarily, but in which it is a linear combination of all the constraints' gradients.

Concretely, suppose we have constraints and are walking along the set of points satisfying Every point on the contour of a given constraint function has a space of allowable directions: the space of vectors perpendicular to The set of directions that are allowed by all constraints is thus the space of directions perpendicular to all of the constraints' gradients. Denote this space of allowable moves by and denote the span of the constraints' gradients by Then the space of vectors perpendicular to every element of

We are still interested in finding points where does not change as we walk, since these points might be (constrained) extrema. We therefore seek such that any allowable direction of movement away from is perpendicular to (otherwise we could increase by moving along that allowable direction). In other words, Thus there are scalars such that

These scalars are the Lagrange multipliers. We now have of them, one for every constraint.

As before, we introduce an auxiliary function

and solve

which amounts to solving equations in unknowns.

The constraint qualification assumption when there are multiple constraints is that the constraint gradients at the relevant point are linearly independent.

Modern formulation via differentiable manifolds

The problem of finding the local maxima and minima subject to constraints can be generalized to finding local maxima and minima on a differentiable manifold [14] In what follows, it is not necessary that be a Euclidean space, or even a Riemannian manifold. All appearances of the gradient (which depends on a choice of Riemannian metric) can be replaced with the exterior derivative

Single constraint

Let be a smooth manifold of dimension Suppose that we wish to find the stationary points of a smooth function when restricted to the submanifold defined by where is a smooth function for which 0 is a regular value.

Let and be the exterior derivatives of and . Stationarity for the restriction at means Equivalently, the kernel contains In other words, and are proportional 1-forms. For this it is necessary and sufficient that the following system of equations holds:

where denotes the exterior product. The stationary points are the solutions of the above system of equations plus the constraint Note that the equations are not independent, since the left-hand side of the equation belongs to the subvariety of consisting of decomposable elements.

In this formulation, it is not necessary to explicitly find the Lagrange multiplier, a number such that

Multiple constraints

Let and be as in the above section regarding the case of a single constraint. Rather than the function described there, now consider a smooth function with component functions for which is a regular value. Let be the submanifold of defined by

is a stationary point of if and only if contains For convenience let and where denotes the tangent map or Jacobian The subspace has dimension smaller than that of , namely and belongs to if and only if belongs to the image of Computationally speaking, the condition is that belongs to the row space of the matrix of or equivalently the column space of the matrix of (the transpose). If denotes the exterior product of the columns of the matrix of the stationary condition for at becomes

Once again, in this formulation it is not necessary to explicitly find the Lagrange multipliers, the numbers such that

Interpretation of the Lagrange multipliers

In this section, we modify the constraint equations from the form to the form where the are m real constants that are considered to be additional arguments of the Lagrangian expression .

Often the Lagrange multipliers have an interpretation as some quantity of interest. For example, by parametrising the constraint's contour line, that is, if the Lagrangian expression is

then

So, λk is the rate of change of the quantity being optimized as a function of the constraint parameter. As examples, in Lagrangian mechanics the equations of motion are derived by finding stationary points of the action, the time integral of the difference between kinetic and potential energy. Thus, the force on a particle due to a scalar potential, F = −∇V, can be interpreted as a Lagrange multiplier determining the change in action (transfer of potential to kinetic energy) following a variation in the particle's constrained trajectory. In control theory this is formulated instead as costate equations.

Moreover, by the envelope theorem the optimal value of a Lagrange multiplier has an interpretation as the marginal effect of the corresponding constraint constant upon the optimal attainable value of the original objective function: If we denote values at the optimum with a star (), then it can be shown that

For example, in economics the optimal profit to a player is calculated subject to a constrained space of actions, where a Lagrange multiplier is the change in the optimal value of the objective function (profit) due to the relaxation of a given constraint (e.g. through a change in income); in such a context is the marginal cost of the constraint, and is referred to as the shadow price. [15]

Sufficient conditions

Sufficient conditions for a constrained local maximum or minimum can be stated in terms of a sequence of principal minors (determinants of upper-left-justified sub-matrices) of the bordered Hessian matrix of second derivatives of the Lagrangian expression. [6] [16]

Examples

Example 1

Illustration of the constrained optimization problem 1 Lagrange very simple.svg
Illustration of the constrained optimization problem 1

Suppose we wish to maximize subject to the constraint The feasible set is the unit circle, and the level sets of f are diagonal lines (with slope −1), so we can see graphically that the maximum occurs at and that the minimum occurs at

For the method of Lagrange multipliers, the constraint is

hence the Lagrangian function,

is a function that is equivalent to when is set to 0.

Now we can calculate the gradient:

and therefore:

Notice that the last equation is the original constraint.

The first two equations yield

By substituting into the last equation we have:

so

which implies that the stationary points of are

Evaluating the objective function f at these points yields

Thus the constrained maximum is and the constrained minimum is .

Example 2

Illustration of the constrained optimization problem 2 Lagrange very simple-1b.svg
Illustration of the constrained optimization problem 2

Now we modify the objective function of Example 1 so that we minimize instead of again along the circle Now the level sets of are still lines of slope −1, and the points on the circle tangent to these level sets are again and These tangency points are maxima of

On the other hand, the minima occur on the level set for (since by its construction cannot take negative values), at and where the level curves of are not tangent to the constraint. The condition that correctly identifies all four points as extrema; the minima are characterized in by and the maxima by

Example 3

Illustration of constrained optimization problem 3. Lagrange simple.svg
Illustration of constrained optimization problem 3.

This example deals with more strenuous calculations, but it is still a single constraint problem.

Suppose one wants to find the maximum values of

with the condition that the - and -coordinates lie on the circle around the origin with radius That is, subject to the constraint

As there is just a single constraint, there is a single multiplier, say

The constraint is identically zero on the circle of radius Any multiple of may be added to leaving unchanged in the region of interest (on the circle where our original constraint is satisfied).

Applying the ordinary Lagrange multiplier method yields

from which the gradient can be calculated:

And therefore:

(iii) is just the original constraint. (i) implies or If then by (iii) and consequently from (ii). If substituting this into (ii) yields Substituting this into (iii) and solving for gives Thus there are six critical points of

Evaluating the objective at these points, one finds that

Therefore, the objective function attains the global maximum (subject to the constraints) at and the global minimum at The point is a local minimum of and is a local maximum of as may be determined by consideration of the Hessian matrix of

Note that while is a critical point of it is not a local extremum of We have

Given any neighbourhood of one can choose a small positive and a small of either sign to get values both greater and less than This can also be seen from the Hessian matrix of evaluated at this point (or indeed at any of the critical points) which is an indefinite matrix. Each of the critical points of is a saddle point of [4]

Example 4

Entropy

Suppose we wish to find the discrete probability distribution on the points with maximal information entropy. This is the same as saying that we wish to find the least structured probability distribution on the points In other words, we wish to maximize the Shannon entropy equation:

For this to be a probability distribution the sum of the probabilities at each point must equal 1, so our constraint is:

We use Lagrange multipliers to find the point of maximum entropy, across all discrete probability distributions on We require that:

which gives a system of n equations, such that:

Carrying out the differentiation of these n equations, we get

This shows that all are equal (because they depend on λ only). By using the constraint

we find

Hence, the uniform distribution is the distribution with the greatest entropy, among distributions on n points.

Example 5

Numerical optimization
Lagrange multipliers cause the critical points to occur at saddle points (Example 5). Lagnum1.png
Lagrange multipliers cause the critical points to occur at saddle points (Example 5).
The magnitude of the gradient can be used to force the critical points to occur at local minima (Example 5). Lagnum2.png
The magnitude of the gradient can be used to force the critical points to occur at local minima (Example 5).

The critical points of Lagrangians occur at saddle points, rather than at local maxima (or minima). [4] [17] Unfortunately, many numerical optimization techniques, such as hill climbing, gradient descent, some of the quasi-Newton methods, among others, are designed to find local maxima (or minima) and not saddle points. For this reason, one must either modify the formulation to ensure that it's a minimization problem (for example, by extremizing the square of the gradient of the Lagrangian as below), or else use an optimization technique that finds stationary points (such as Newton's method without an extremum seeking line search) and not necessarily extrema.

As a simple example, consider the problem of finding the value of x that minimizes constrained such that (This problem is somewhat untypical because there are only two values that satisfy this constraint, but it is useful for illustration purposes because the corresponding unconstrained function can be visualized in three dimensions.)

Using Lagrange multipliers, this problem can be converted into an unconstrained optimization problem:

The two critical points occur at saddle points where x = 1 and x = −1.

In order to solve this problem with a numerical optimization technique, we must first transform this problem such that the critical points occur at local minima. This is done by computing the magnitude of the gradient of the unconstrained optimization problem.

First, we compute the partial derivative of the unconstrained problem with respect to each variable:

If the target function is not easily differentiable, the differential with respect to each variable can be approximated as

where is a small value.

Next, we compute the magnitude of the gradient, which is the square root of the sum of the squares of the partial derivatives:

(Since magnitude is always non-negative, optimizing over the squared-magnitude is equivalent to optimizing over the magnitude. Thus, the "square root" may be omitted from these equations with no expected difference in the results of optimization.)

The critical points of h occur at x = 1 and x = −1, just as in Unlike the critical points in however, the critical points in h occur at local minima, so numerical optimization techniques can be used to find them.

Applications

Control theory

In optimal control theory, the Lagrange multipliers are interpreted as costate variables, and Lagrange multipliers are reformulated as the minimization of the Hamiltonian, in Pontryagin's minimum principle.

Nonlinear programming

The Lagrange multiplier method has several generalizations. In nonlinear programming there are several multiplier rules, e.g. the Carathéodory–John Multiplier Rule and the Convex Multiplier Rule, for inequality constraints. [18]

Power systems

Methods based on Lagrange multipliers have applications in power systems, e.g. in distributed-energy-resources (DER) placement and load shedding. [19]

Safe Reinforcement Learning

The method of Lagrange multipliers applies to constrained Markov decision processes. [20] It naturally produces gradient-based primal-dual algorithms in safe reinforcement learning. [21]

See also

Related Research Articles

In statistics, maximum likelihood estimation (MLE) is a method of estimating the parameters of an assumed probability distribution, given some observed data. This is achieved by maximizing a likelihood function so that, under the assumed statistical model, the observed data is most probable. The point in the parameter space that maximizes the likelihood function is called the maximum likelihood estimate. The logic of maximum likelihood is both intuitive and flexible, and as such the method has become a dominant means of statistical inference.

The calculus of variations is a field of mathematical analysis that uses variations, which are small changes in functions and functionals, to find maxima and minima of functionals: mappings from a set of functions to the real numbers. Functionals are often expressed as definite integrals involving functions and their derivatives. Functions that maximize or minimize functionals may be found using the Euler–Lagrange equation of the calculus of variations.

In mathematics, the Hessian matrix, Hessian or Hesse matrix is a square matrix of second-order partial derivatives of a scalar-valued function, or scalar field. It describes the local curvature of a function of many variables. The Hessian matrix was developed in the 19th century by the German mathematician Ludwig Otto Hesse and later named after him. Hesse originally used the term "functional determinants". The Hessian is sometimes denoted by H or, ambiguously, by ∇2.

Pontryagin's maximum principle is used in optimal control theory to find the best possible control for taking a dynamical system from one state to another, especially in the presence of constraints for the state or input controls. It states that it is necessary for any optimal control along with the optimal state trajectory to solve the so-called Hamiltonian system, which is a two-point boundary value problem, plus a maximum condition of the control Hamiltonian. These necessary conditions become sufficient under certain convexity conditions on the objective and constraint functions.

In physics, a first class constraint is a dynamical quantity in a constrained Hamiltonian system whose Poisson bracket with all the other constraints vanishes on the constraint surface in phase space. To calculate the first class constraint, one assumes that there are no second class constraints, or that they have been calculated previously, and their Dirac brackets generated.

<span class="mw-page-title-main">Interior-point method</span> Algorithms for solving convex optimization problems

Interior-point methods are algorithms for solving linear and non-linear convex optimization problems. IPMs combine two advantages of previously-known algorithms:

Convex optimization is a subfield of mathematical optimization that studies the problem of minimizing convex functions over convex sets. Many classes of convex optimization problems admit polynomial-time algorithms, whereas mathematical optimization is in general NP-hard.

In statistics and information theory, a maximum entropy probability distribution has entropy that is at least as great as that of all other members of a specified class of probability distributions. According to the principle of maximum entropy, if nothing is known about a distribution except that it belongs to a certain class, then the distribution with the largest entropy should be chosen as the least-informative default. The motivation is twofold: first, maximizing entropy minimizes the amount of prior information built into the distribution; second, many physical systems tend to move towards maximal entropy configurations over time.

In mathematical optimization, the Karush–Kuhn–Tucker (KKT) conditions, also known as the Kuhn–Tucker conditions, are first derivative tests for a solution in nonlinear programming to be optimal, provided that some regularity conditions are satisfied.

In mathematical optimization theory, duality or the duality principle is the principle that optimization problems may be viewed from either of two perspectives, the primal problem or the dual problem. If the primal is a minimization problem then the dual is a maximization problem. Any feasible solution to the primal (minimization) problem is at least as large as any feasible solution to the dual (maximization) problem. Therefore, the solution to the primal is an upper bound to the solution of the dual, and the solution of the dual is a lower bound to the solution of the primal. This fact is called weak duality.

In mathematics and economics, the envelope theorem is a major result about the differentiability properties of the value function of a parameterized optimization problem. As we change parameters of the objective, the envelope theorem shows that, in a certain sense, changes in the optimizer of the objective do not contribute to the change in the objective function. The envelope theorem is an important tool for comparative statics of optimization models.

Sequential quadratic programming (SQP) is an iterative method for constrained nonlinear optimization which may be considered a quasi-Newton method. SQP methods are used on mathematical problems for which the objective function and the constraints are twice continuously differentiable, but not necessarily convex.

Covariance matrix adaptation evolution strategy (CMA-ES) is a particular kind of strategy for numerical optimization. Evolution strategies (ES) are stochastic, derivative-free methods for numerical optimization of non-linear or non-convex continuous optimization problems. They belong to the class of evolutionary algorithms and evolutionary computation. An evolutionary algorithm is broadly based on the principle of biological evolution, namely the repeated interplay of variation and selection: in each generation (iteration) new individuals are generated by variation, usually in a stochastic way, of the current parental individuals. Then, some individuals are selected to become the parents in the next generation based on their fitness or objective function value . Like this, over the generation sequence, individuals with better and better -values are generated.

The adjoint state method is a numerical method for efficiently computing the gradient of a function or operator in a numerical optimization problem. It has applications in geophysics, seismic imaging, photonics and more recently in neural networks.

<span class="mw-page-title-main">Total variation denoising</span> Noise removal process during image processing

In signal processing, particularly image processing, total variation denoising, also known as total variation regularization or total variation filtering, is a noise removal process (filter). It is based on the principle that signals with excessive and possibly spurious detail have high total variation, that is, the integral of the image gradient magnitude is high. According to this principle, reducing the total variation of the signal—subject to it being a close match to the original signal—removes unwanted detail whilst preserving important details such as edges. The concept was pioneered by L. I. Rudin, S. Osher, and E. Fatemi in 1992 and so is today known as the ROF model.

Augmented Lagrangian methods are a certain class of algorithms for solving constrained optimization problems. They have similarities to penalty methods in that they replace a constrained optimization problem by a series of unconstrained problems and add a penalty term to the objective, but the augmented Lagrangian method adds yet another term designed to mimic a Lagrange multiplier. The augmented Lagrangian is related to, but not identical with, the method of Lagrange multipliers.

In mathematical physics, the Belinfante–Rosenfeld tensor is a modification of the stress–energy tensor that is constructed from the canonical stress–energy tensor and the spin current so as to be symmetric yet still conserved.

Lagrangian field theory is a formalism in classical field theory. It is the field-theoretic analogue of Lagrangian mechanics. Lagrangian mechanics is used to analyze the motion of a system of discrete particles each with a finite number of degrees of freedom. Lagrangian field theory applies to continua and fields, which have an infinite number of degrees of freedom.

Proximal gradientmethods for learning is an area of research in optimization and statistical learning theory which studies algorithms for a general class of convex regularization problems where the regularization penalty may not be differentiable. One such example is regularization of the form

In mathematical optimization, the proximal operator is an operator associated with a proper, lower semi-continuous convex function from a Hilbert space to , and is defined by:

References

  1. Hoffmann, Laurence D.; Bradley, Gerald L. (2004). Calculus for Business, Economics, and the Social and Life Sciences (8th ed.). pp. 575–588. ISBN   0-07-242432-X.
  2. Beavis, Brian; Dobbs, Ian M. (1990). "Static Optimization". Optimization and Stability Theory for Economic Analysis. New York: Cambridge University Press. p. 40. ISBN   0-521-33605-8.
  3. Protter, Murray H.; Morrey, Charles B. Jr. (1985). Intermediate Calculus (2nd ed.). New York, NY: Springer. p. 267. ISBN   0-387-96058-9.
  4. 1 2 3 Walsh, G.R. (1975). "Saddle-point Property of Lagrangian Function". Methods of Optimization. New York, NY: John Wiley & Sons. pp. 39–44. ISBN   0-471-91922-5.
  5. Kalman, Dan (2009). "Leveling with Lagrange: An alternate view of constrained optimization". Mathematics Magazine . 82 (3): 186–196. doi:10.1080/0025570X.2009.11953617. JSTOR   27765899. S2CID   121070192.
  6. 1 2 Silberberg, Eugene; Suen, Wing (2001). The Structure of Economics: A Mathematical Analysis (Third ed.). Boston: Irwin McGraw-Hill. pp. 134–141. ISBN   0-07-234352-4.
  7. de la Fuente, Angel (2000). Mathematical Methods and Models for Economists . Cambridge: Cambridge University Press. p.  285. doi:10.1017/CBO9780511810756. ISBN   978-0-521-58512-5.
  8. Luenberger, David G. (1969). Optimization by Vector Space Methods. New York: John Wiley & Sons. pp. 188–189.
  9. Bertsekas, Dimitri P. (1999). Nonlinear Programming (Second ed.). Cambridge, MA: Athena Scientific. ISBN   1-886529-00-0.
  10. Vapnyarskii, I.B. (2001) [1994], "Lagrange multipliers", Encyclopedia of Mathematics , EMS Press .
  11. Lasdon, Leon S. (2002) [1970]. Optimization Theory for Large Systems (reprint ed.). Mineola, New York, NY: Dover. ISBN   0-486-41999-1. MR   1888251.
  12. Hiriart-Urruty, Jean-Baptiste; Lemaréchal, Claude (1993). "Chapter XII: Abstract duality for practitioners". Convex analysis and minimization algorithms. Grundlehren der Mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences]. Vol. 306. Berlin, DE: Springer-Verlag. pp. 136–193 (and Bibliographical comments pp. 334–335). ISBN   3-540-56852-2. MR   1295240. Volume II: Advanced theory and bundle methods.
  13. Lemaréchal, Claude (15–19 May 2000). "Lagrangian relaxation". In Jünger, Michael; Naddef, Denis (eds.). Computational combinatorial optimization: Papers from the Spring School held in Schloß Dagstuhl. Spring School held in Schloß Dagstuhl, May 15–19, 2000. Lecture Notes in Computer Science. Vol. 2241. Berlin, DE: Springer-Verlag (published 2001). pp. 112–156. doi:10.1007/3-540-45586-8_4. ISBN   3-540-42877-1. MR   1900016. S2CID   9048698.
  14. Lafontaine, Jacques (2015). An Introduction to Differential Manifolds. Springer. p. 70. ISBN   978-3-319-20735-3.
  15. Dixit, Avinash K. (1990). "Shadow Prices". Optimization in Economic Theory (2nd ed.). New York: Oxford University Press. pp. 40–54. ISBN   0-19-877210-6.
  16. Chiang, Alpha C. (1984). Fundamental Methods of Mathematical Economics (Third ed.). McGraw-Hill. p.  386. ISBN   0-07-010813-7.
  17. Heath, Michael T. (2005). Scientific Computing: An introductory survey. McGraw-Hill. p. 203. ISBN   978-0-07-124489-3.
  18. Pourciau, Bruce H. (1980). "Modern multiplier rules". American Mathematical Monthly . 87 (6): 433–452. doi:10.2307/2320250. JSTOR   2320250.
  19. Gautam, Mukesh; Bhusal, Narayan; Benidris, Mohammed (2020). A sensitivity-based approach to adaptive under-frequency load shedding. 2020 IEEE Texas Power and Energy Conference (TPEC). Institute of Electronic and Electrical Engineers. pp. 1–5. doi:10.1109/TPEC48276.2020.9042569.
  20. Altman, Eitan (2021). Constrained Markov Decision Processes. Routledge.
  21. Ding, Dongsheng; Zhang, Kaiqing; Jovanovic, Mihailo; Basar, Tamer (2020). Natural policy gradient primal-dual method for constrained Markov decision processes. Advances in Neural Information Processing Systems.

Further reading

Exposition
Additional text and interactive applets