Smooth maps minimizing the energy and the calibrated geometry (Q6084647): Difference between revisions

The theory of calibrated geometry sees new progression with this article. Let \(X, Y\) be smooth manifolds of dimension \(m\) and \(n\), respectively. Let \(X\) be compact and oriented. A volume form \(\mathrm{vol}\in\Omega^m(X)\) is assumed to exist, a nowhere vanishing \(m\)-form which determines an orientation and a measure on \(X\). Given also a map \(\sigma:C^\infty(X , Y )\rightarrow L^1 (X )\), the author defines the energy map \(\mathcal{E}:C^\infty(X , Y ) \rightarrow \mathbb{R}\) as \[ \mathcal{E}(f) :=\int_X\sigma(f)\,\mathrm{vol}. \] The article is concerned with the lower bound and the minimum of \(\mathcal{E}\) restricted to the set of functions in a given homotopy class of any chosen function \(f\). For this purpose, it recurs to the following concepts strongly related with the theory mentioned in the second part of the title. First, denoting the graph function by \((1_X,f)\), meaning the function \(X\rightarrow X\times Y\) defined by \((x,f(x)),\ x\in X\), there comes the relevant notion: \(\Phi\in\Omega^m(X \times Y )\) is a \(\sigma\)-calibration if \(\mathrm{d}\phi=0\) and \((1_X , f )^*\Phi \leq \sigma( f )\mathrm{vol}\) for any smooth map \(f : X \rightarrow Y\). Moreover, \(f\) is a \((\sigma,\Phi)\)-calibrated map if \((1_X , f )^*\Phi = \sigma ( f )\mathrm{vol}\). The expected properties of these concepts follow, namely invariance of the integral in the homotopy class and the fact that any \((\sigma,\Phi)\)-calibrated map minimizes \(\mathcal{E}\) in its homotopy class. A typical energy density map is when \(X,Y\) are Riemannian and \(\sigma_k(f):=|\mathrm{d}f|^k\). In this case the energy map is denoted by \(\mathcal{E}_k\). Another energy density is given by \(\sigma_{p,q}(f)=|\mathrm{d}f|^q_p\), where the norm is essentially the Hilbert-Schmidt norm. A main result is the following: For any \(1\leq k \leq m=\dim X\), there exists a positive constant \(C\) depending only on \(k\), \((X , g)\), \((Y , h)\) (Riemannian) and the basis of \(H^k(X ,\mathbb{R})\), \(H^k_{\mathrm{bdd}}(Y,\mathbb{R})\) such that for any \(f \in C^\infty(X , Y )\) we have \[\mathcal{E}_k( f ) \geq C|P|,\] where \(P\) is the matrix of \([f^*]^k_{\mathrm{bdd}}:H^k_{\mathrm{bdd}}(Y,\mathbb{R})\rightarrow H^k(X,\mathbb{R})\). In particular, if \([f^*]^k_{\mathrm{bdd}}\) is nonzero, then \(\inf({\mathcal{E}_k}_{|[ f ]})\) is positive. Several applications are given for manifolds of special holonomy, namely for Kähler manifolds, quaternionic Kähler, \(\mathrm{G}_2\) and \(\mathrm{Spin}(7)\), finding old and new results regarding the homotopy class of the identity of \(X\). The calibration \(\Phi\) follows from a canonical calibration, a \(k\)-form \(\varphi\), on the special holonomy \((X,g,\varphi)\) (one wonders if the condition on the holonomy being strict or the calibration being harmonic could be dropped to the case of \(\varphi\) just a closed form). The main conclusion is that the map \(1_X\) is a \((\sigma_{k,k},\Phi)\)-calibrated map. \textit{C. Croke} and \textit{A. Fathi} [Bull. Lond. Math. Soc. 22, No. 5, 489--494 (1990; Zbl 0719.53020)] introduced a homotopy invariant of a smooth map \(f : X \rightarrow Y\) which gives the lower bound to the 2-energy \(\mathcal{E}_2\). In the last section of the present article, a new invariant more efficient than that one is introduced.