Math 2.0

Calculus of variations and optimal control; optimization Definition

Direct method in the calculus of variations

Builds on Calculus of variationsReflexive spaceWeak topologySemi-continuity

Definition

Let XX be a set with a notion of convergence of sequences, and let J ⁣:X→ℝ∪{+∞}J \colon X \to \R \cup \set{+\infty} be a functional with inf⁡XJ<+∞\inf_X J < +\infty. The direct method is the following way to prove that JJ has a minimiser.

  1. Take a minimising sequence (uk)(u_k), that is, J(uk)→inf⁡XJJ(u_k) \to \inf_X J.
  2. Use coercivity to show that (uk)(u_k) lies in a set that is sequentially compact, and extract a subsequence ukj→uu_{k_j} \to u in XX.
  3. Use sequential lower semicontinuity, J(u)≤lim inf⁡jJ(ukj)J(u) \le \liminf_j J(u_{k_j}), to conclude that J(u)=inf⁡XJJ(u) = \inf_X J.

The standard abstract form is this. Let XX be a reflexive Banach space and let JJ be coercive, meaning J(u)→+∞J(u) \to +\infty as ∥u∥→∞\norm{u} \to \infty. If JJ is sequentially lower semicontinuous for the weak topology, then JJ attains its minimum on XX.

Remarks

In the abstract form, coercivity makes a minimising sequence bounded, and bounded sequences in a reflexive Banach space have weakly convergent subsequences. For integral functionals J(u)=∫ΩL(x,u,∇u) dxJ(u) = \int_\Omega L(x, u, \nabla u) \, dx of the calculus of variations on W1,q(Ω)W^{1,q}(\Omega) with 1<q<∞1 < q < \infty, weak lower semicontinuity holds when LL is continuous, bounded below and convex in ∇u\nabla u.

Example: let Ω⊆ℝn\Omega \subseteq \R^n be bounded and open, f∈L2(Ω)f \in L^2(\Omega), and J(u)=∫Ω(12∣∇u∣2−fu) dxJ(u) = \int_\Omega \big(\frac{1}{2}\abs{\nabla u}^2 - f u\big) \, dx on H01(Ω)H^1_0(\Omega). By the Poincaré inequality J(u)≥c∥u∥H12−C∥u∥H1J(u) \ge c\norm{u}_{H^1}^2 - C\norm{u}_{H^1}, so JJ is coercive; it is convex and continuous, hence weakly lower semicontinuous. So JJ has a minimiser, which is the weak solution of −Δu=f-\Delta u = f with u=0u = 0 on ∂Ω\partial\Omega.

Non-example: J(u)=∫01((u′2−1)2+u2) dxJ(u) = \int_0^1 \big((u'^2 - 1)^2 + u^2\big) \, dx on W1,4(0,1)W^{1,4}(0, 1) is coercive, but its integrand is not convex in u′u', so it is not weakly lower semicontinuous. Indeed inf⁡J=0\inf J = 0, approached by sawtooth functions with slopes ±1\pm 1 and small height, but no function has J(u)=0J(u) = 0, since this would need u=0u = 0 and ∣u′∣=1\abs{u'} = 1 almost everywhere.

Reading path

To understand this page from the ground up, read these 5 pages in order. The foundations come first.

Show the path as a graph
Each page uses the pages above it. The colour of the bar on a box shows how strong the evidence is, and the box also gives it as text. The list has the same pages.
  • no evidence level
  1. Calculus of variations concept
  2. Reflexive space concept
  3. Semi-continuity concept
  4. Weak topology concept
  5. Direct method in the calculus of variations definition

Further reading

  1. Web https://en.wikipedia.org/wiki/Direct_method_in_the_calculus_of_variations
About this page
TagVAJPHCQBHR
URIm2://VAJPHCQBHR
Kinddefinition
WritingW0 stub
MSC49J10, 49J45, 35A15
DefinesDirect method in the calculus of variations
UsesCalculus of variations, Reflexive space, Weak topology, Semi-continuity

Markdown · JSON · Edit on GitHub

What this page needs for W1

  • an inventory row: the row of the inventory that the node accounts for (0 of 1)
  • a source with a locator (0 of 1)

The writing levels say how complete the text of a page is. They do not measure its truth.

What links here

No page links to this page yet.