Calculus of variations is a branch of mathematical analysis that deals with finding functions that optimize (maximize or minimize) functionals. Lower semicontinuity is a fundamental concept in this field, playing a crucial role in ensuring the existence of minimizers for variational problems involving multiple integrals. This article explores the concept, its mathematical foundations, and its significance in the calculus of variations.
In the calculus of variations, we often encounter functionals of the form:
where ℝn is an open domain, u: ℝm is our function (or function vector), and f: ℝm ℝmn ℝ is the integrand. The goal is to find functions u that minimize this functional under certain constraints.
For such problems, the existence of minimizers hinges on two main properties: coercivity and lower semicontinuity of the functional I. While coercivity ensures that the functional is bounded from below, lower semicontinuity provides the necessary topological structure to guarantee that minimizing sequences converge to an actual minimizer.
Definition: A functional I: X [-, +], where X is a topological space, is said to be lower semicontinuous at x0 X if for every > 0, there exists a neighborhood U of x0 such that for all x U, I(x) > I(x0) - .
Equivalently, I is lower semicontinuous at x0 if for any sequence {xk} converging to x0, we have I(x0) liminf I(xk). The functional is called lower semicontinuous if it has this property at every point in its domain.
In the context of variational problems with multiple integrals, we typically work with the topology induced by Lp norms or weak topologies in various Sobolev spaces. The choice of topology is crucial, as a functional might be lower semicontinuous in one topology but not in another.
The direct method of the calculus of variations is a powerful technique for proving the existence of minimizers for variational problems. This method relies heavily on lower semicontinuity. The direct method consists of three main steps:
Without lower semicontinuity, even if a minimizing sequence converges to some limit, the value of the functional at that limit might be greater than the infimum of the functional values, meaning the limit is not a true minimizer. This is why establishing lower semicontinuity is essential for proving the existence of solutions in variational problems.
One of the most important results connecting the geometry of the integrand with lower semicontinuity of the associated functional is the following theorem:
Theorem: If the integrand f(x,s,p) is convex in p (the gradient variable) and satisfies certain measurability and growth conditions, then the functional I is sequentially weakly lower semicontinuous in the appropriate Sobolev space.
This theorem essentially tells us that convexity of the integrand with respect to the gradient variable is a sufficient condition for lower semicontinuity of the functional. This is why many problems in physics and mechanics, which often involve convex energy densities, have well-posed variational formulations.
For multiple integrals, the notion of convexity needs to be refined. The appropriate concept for problems in higher dimensions is quasiconvexity, introduced by Morrey:
Definition: A function f: ℝmn ℝ is quasiconvex if for every matrix A ℝmn and every W01,(D; ℝm), where D ℝn is a bounded domain, we have:
Quasiconvexity plays a role analogous to convexity in one-dimensional problems. The following theorem establishes its connection to lower semicontinuity:
Theorem: If the integrand f(x,s,p) is measurable in x, continuous in s, and quasiconvex in p, and satisfies appropriate growth conditions, then the functional I is sequentially weakly lower semicontinuous.
This result is fundamental in the modern calculus of variations, as it provides necessary and almost sufficient conditions for lower semicontinuity of multiple integrals in the vectorial case.
When the integrand is not convex (or quasiconvex in the multi-dimensional case), the functional may fail to be lower semicontinuous. In such situations, the direct method of the calculus of variations breaks down, and minimizers may not exist. This leads to the concept of relaxation.
Relaxation involves finding the lower semicontinuous envelope of the functional, defined as:
The relaxed functional Isc is lower semicontinuous by construction, and its minimizers (which exist under suitable conditions) can be interpreted as "generalized" solutions to the original problem. The theory of relaxation has important applications in elasticity theory and materials science, where many energy functionals are non-convex.
For variational problems defined on Sobolev spaces, different notions of convergence yield different requirements for lower semicontinuity. In particular:
This distinction is crucial because compactness results in Sobolev spaces typically provide weak convergence, not strong convergence. Therefore, for the direct method to be applicable, we need to establish weak lower semicontinuity, which necessitates the convexity or quasiconvexity conditions mentioned earlier.
The theory of lower semicontinuity of multiple integrals has numerous applications in physics, engineering, and economics:
Lower semicontinuity of multiple integrals is a cornerstone concept in the calculus of variations. It provides the necessary conditions for the existence of minimizers, which is fundamental to solving many variational problems arising in mathematics and applied sciences. The connection between convexity/quasiconvexity of the integrand and lower semicontinuity of the functional represents one of the most fruitful areas of research in this field.
Understanding this concept allows mathematicians and scientists to determine when variational problems have well-posed solutions and provides tools for handling situations where standard existence theorems fail. The theory continues to evolve, with ongoing research extending these concepts to more general settings and exploring applications in emerging fields.
