Noncontact set 2026-10-31
The noncontact set is where the unknown function is strictly above its obstacle. For continuous functions it is open. Every smooth variation with compact support in this set remains feasible with either sign when its amplitude is small enough. The first variation therefore vanishes there, yielding the unconstrained Euler-Lagrange equation. On the complementary contact set of an obstacle problem, only the appropriate one-sided variational inequality is available globally.
Obstacle problem Created 2026-10-06 Updated 2026-11-05
A classical obstacle problem minimizes the Dirichlet energy with forcing while requiring the unknown function to stay above an obstacle. On a bounded Lipschitz domain , take and an obstacle . Minimize
The closed convex set is nonempty, and the Poincaré inequality makes coercive and strictly convex. There is a unique minimizer , characterized by the variational inequality
It follows by differentiating at ; convexity also makes the inequality sufficient for minimality.
Where regularity permits pointwise notation, the same conditions read , , and . On the noncontact set, variations of both signs recover the unconstrained Poisson equation. On the contact set of an obstacle problem, the obstacle can supply a reaction force. Its interface is an unknown free boundary problem. An analogous separation into stopping and continuation regions appears in optimal stopping, where the value must dominate the immediate payoff.
For a string of positive tension with fixed endpoints , zero transverse loading and obstacle , , equilibrium minimizes among functions above . On a free interval , while in contact . Smooth tangency of the right line through to the parabola at requires , yielding the displayed root in .
The two pieces join with equal value and slope. On the free portions ; in contact is the nonnegative reaction. These facts verify the complementarity conditions. The associated variational inequality and the exact energy difference prove global minimality; strict convexity with fixed endpoints proves uniqueness. At contact reduces to one point; at it fills the whole interval.
For any admissible , put . At endpoints . On the contact set, and imply , while away from contact the reaction is zero. Hence everywhere. Integration by parts gives
Rearranging proves the requested variational inequality. Piecewise integration is legitimate because the slope is continuous at the contact boundary, so there are no omitted point-force terms.
The small-slope excess length is , so the first term is the work needed against uniform tension. The second is the potential energy of the downward distributed load. Their sum is total potential energy subject to the obstacle constraint.
Expand the difference exactly:
The first two integrals together are nonnegative by part (b), and the square is nonnegative. Thus
For this also proves uniqueness: equality for two minimizers forces identical derivatives, and the fixed endpoints remove their additive constant. The convex admissible set and strictly convex energy make the variational inequality a global minimization condition.
For any , the segment stays in for . Since is a minimizer, the right derivative of its energy at is nonnegative. The first variation therefore gives the variational inequality
The factor two comes from the quadratic gradient term. Only a one-sided inequality follows because a variation pointing below the obstacle need not be admissible.
The noncontact set is open. For , the positive gap above the obstacle has a positive minimum on the support of . Thus both and are admissible for sufficiently small . Applying the variational inequality to both signs yields
Hence the Euler-Lagrange equation on the noncontact set is
It holds classically under the assumed smoothness. On the contact set the obstacle prevents the negative variations used to obtain this equality, so a variational inequality remains the appropriate global description.
Let both solve the variational inequality. Test the inequality for with and the one for with , then add. Writing gives
since cosine is Lipschitz with constant one. Since , the fundamental theorem of calculus and Cauchy-Schwarz inequality give ; integrating gives . Therefore
For , the coefficient is positive, so and the zero boundary value gives . This proves uniqueness with an explicit sufficient smallness condition. More generally any Poincaré inequality constant gives the condition ; optimizing that constant is unnecessary for the requested conclusion. The underlying principle is uniqueness for a strongly monotone variational inequality.
A useful form of Minty's lemma is the following. Let be a real Banach space, a nonempty convex set, and a monotone operator that is hemicontinuous. Here operator monotonicity means for , and hemicontinuity means that is continuous on for and . For ,
The forward implication follows from operator monotonicity:
For the reverse implication, fix and take for . The assumed inequality gives
Divide by and send . Hemicontinuity gives , as required.
In the whole-space version, for a monotone operator that is hemicontinuous,
Apply the just-proved result to and ; the resulting inequality for all forces by choosing . This is the variational inequality form of Minty's lemma.
Take a finite set of source-sink pairs , with fixed demands , and finitely many permitted routes for each pair. Every positive demand must have at least one available route. Let be route flow and link throughput. The source-sink route incidence matrix and link-route incidence matrix are
for simple routes; if a route traverses a link repeatedly, its entry counts those traversals. Thus conserves each demand and aggregates link traffic.
At a Wardrop equilibrium, every route carrying positive flow has minimum delay among routes for the same source-sink pair. If
there are numbers such that for , with equality whenever . This is a nonatomic routing game: one infinitesimal user's route change does not alter the link costs it sees.
For continuous nondecreasing link delays, define the Beckmann potential
Each term is convex because its derivative is nondecreasing. The feasible route-flow set is nonempty and compact: on each pair's routes. Hence the continuous function attains a minimum. Its gradient with respect to route flow is
If carries positive flow and serves the same pair, moving a small amount of flow from to is feasible. At a minimizer its directional derivative is nonnegative. Thus every used route minimizes delay, proving existence of a Wardrop equilibrium.
Conversely, at such an equilibrium, any feasible competitor satisfies
Indeed, the competitor's cost sum for pair is at least , while the current flow's cost sum equals . This first-order variational inequality is sufficient for optimality of a differentiable convex function. Therefore
The argument proves both directions of the claimed optimization equivalence, including unused routes with costs tied at the minimum.
If each is strictly increasing, each integral is strictly convex, hence so is as a function of link throughputs. The feasible link-throughput set is convex. Two distinct minimizing vectors would have a feasible midpoint with strictly smaller potential, so the equilibrium link-throughput vector is unique. Strictly convex function in need not give strictly convex function in , because can have a nontrivial kernel.
For an explicit route-flow ambiguity, put two parallel links in each of two consecutive stages, with one unit of demand traversing both stages. Order the four routes by the first and second link choices: . Set all four link delays to . Then
induces throughput on each link, and every route has delay one. Every member of this family is an equilibrium, though its route-flow vector changes. This is route-flow nonuniqueness at a Wardrop equilibrium with strictly increasing delays.
For Braess paradox, use vertices and one unit of -to- demand. The delays are , , , and . Initially the only routes are and . If the upper route carries , their costs are and , so equilibrium has and common delay .
Add the directed link with delay zero. Let the upper, lower and new middle routes carry , where . Their costs become
If , the middle route is strictly cheaper than the upper route, contradicting equilibrium; likewise is impossible. Thus all flow uses the middle route, with , and all three route costs are two. Therefore
Figure 1.
Adding the central route raises the selfish equilibrium delay in a four-node network
. Details.
The potential for the expanded network is
The old flow remains feasible and has potential ; the new equilibrium has potential . The enlarged feasible set has indeed lowered the minimized potential. But total travel delay is , not this integral potential: it rises from to . This explains the paradox within the optimization framework. The optimal total-delay value cannot worsen when the old flow is still allowed; selfish equilibrium minimizes a different function.
The phenomenon also survives strictly increasing delays on every link. Replace each constant-one delay by and the new-link delay by . Symmetry gives upper and lower flows after expansion, with middle flow . The common new delay is , while the old delay was . For , these are . Thus the example does not depend on interpreting increasing delays as merely nondecreasing.
For a steady flow network, let each origin-destination class have fixed demand . Route flows satisfy . The link flow is , and a link has continuous travel delay . Its route delay is .
A Wardrop equilibrium describes nonatomic traffic: each traveler is too small to alter aggregate delays by changing route. For each class there is a minimum delay such that
Thus all used routes of one class have equal minimum cost, and an unused route cannot offer a shorter trip. The condition concerns private travel time, rather than total network delay.
Equivalently, for every feasible route vector ,
Indeed, each used route has class minimum cost, so reallocating demand cannot reduce the cost evaluated at the original flows; conversely a positive flow on a non-minimum route gives an improving transfer. This variational inequality form remains meaningful even when route costs are not separable link functions.
For the log-cosh obstacle variational inequality, the admissible set is convex: for any , remains above the obstacle and in for . Minimality and the right derivative at zero therefore give
The one-sided sign comes from admissibility, which permits motion toward every but does not permit an arbitrary negative variation at a contact point.
Now fix . Smoothness on and continuity of the obstacle give a strictly positive minimum of on the compact support of . Hence is admissible for both signs of sufficiently small . Applying the inequality to both variations yields
Integration by parts gives distributionally on , and smoothness makes it pointwise. Thus
On the contact set the variational inequality, rather than unrestricted stationarity, remains the governing statement. Convexity of the energy also makes the inequality sufficient for minimization; strict monotonicity implies at most one minimizer, although existence was already stipulated in this part.
Unilateral contact complementarity Created 2026-10-31 Updated 2026-11-05
For a nonadhesive obstacle, is the signed separation and its outward normal contact force. Nonpenetration gives , absence of adhesion gives , and reaction is possible only at contact, giving . In a small-slope taut-string model, and force balance gives for downward applied load . The three conditions then become , , . Multiplication by admissible displacement differences and integration gives the associated variational inequality, connecting a local contact law to constrained energy minimization.
Let be a convex set in a real Banach space, and let be a strongly monotone operator. If both satisfy the displayed variational inequality, testing with and respectively and adding yields . Strong monotonicity gives , hence . This proves uniqueness whenever a solution exists.
Variational inequality 2026-10-07
For a feasible set and a vector field , a variational inequality asks for satisfying the displayed inequality for every feasible . If is convex and is the gradient of a differentiable convex function, it is the necessary and sufficient first-order condition for minimizing that function. It also describes Wardrop equilibria when no scalar objective has the route-cost vector as its gradient.