Transformation de Fourier et majoration de sommes exponentielles. (Fourier transformation and majoration of exponential sums.). (French) Zbl 0603.14015

Katz, Nicholas M.; Laumon, Gérard

Publ. Math., Inst. Hautes Étud. Sci. 62, 145-202 (1985).

For a finite field \(k={\mathbb{F}}_ q\) of characteristic \(p\) one has the Lang torsor \((=\) principal fibre bundle) \({\mathcal L}\) on the affine line \({\mathbb{A}}^ 1_ k=Spec k[U]\), defined by the equation \(V^ q-V=U\) and for \(\ell \neq p\) the \({\bar {\mathbb{Q}}}^{\times}_{\ell}\)-valued additive characters \(\psi\) of \({\mathbb{F}}_ q\) give a set of \({\bar {\mathbb{Q}}}^{\times}_{\ell}\)-bundles \({\mathcal L}_{\psi}\) (Artin-Schreier sheaves) on the line. Let E be the \({\mathbb{F}}_ q\)-vector space \({\mathbb{F}}^ r_ q\) of rank r and \(E^{\vee}\) its dual. The category D(E,\({\bar {\mathbb{Q}}}_{\ell})\) is the derived category of \({\bar {\mathbb{Q}}}_{\ell}\)-sheaves on E (bounded complexes with negative degrees only and residual cohomology with constructible supports). One then has the \(\ell\)-adic Fourier transformation \({\mathcal F}: D(E,{\bar {\mathbb{Q}}}_{\ell})\to D(E^{\vee},{\bar {\mathbb{Q}}}_{\ell}):\) the definition formula for the elementary Fourier transform \({\mathcal F}(K)=<K,\psi >\) now reads like \({\mathcal F}(K)=Rpr_*^{\vee}(pr^*K^ L\otimes \mu^*{\mathcal L}_{\psi})[r],\) where \(\mu\) is the evaluation map \(E^{\vee}\times E\to {\mathbb{A}}^ 1\). Actually E may, more generally, be a vector bundle on a good \({\mathbb{F}}_ q\)-scheme. \({\mathcal F}\) commutes with duality. One proves a Riemann-Roch theorem for objects K adapted to a given stratification of E, relating the Euler-Poincaré characteristics \(\chi(F(K))\) and \(\chi(K)\) which are defined as functions on \(E^{\vee}\) and E. If in the definition formula for \(\chi\) (K) the signs \((-)^ i\) are all replaced by \(+\), one gets a function on E, denoted by \(\| K\|\). One of the main results is a uniformity theorem: \((\sup \| {\mathcal F}(K)\|)\leq \text{const. }(\sup \| K\|)\), where the suprema are taken along a fibre of \(E^{\vee}\) and E, with an integer constant determined by E (which is allowed to be given as a vector bundle over a \({\mathbb{Z}}\)-scheme of finite type). The uniformity theorem is applied in the following situation. Let R be a subring of \({\mathbb{C}}\) of finite type, X an affine R-scheme of relative dimension \(m\) and \(\phi: R\to k\) a ring homomorphism. Let \(f: X\to {\mathbb{A}}^ r_ R\) be an R- morphism and g an invertible rational function on X. Then, with \(f_{\phi}: X_{\phi}\to {\mathbb{A}}^ r_ R\) and \(g_{\phi}: X_{\phi}\to {\mathbb{G}}_ m={\mathbb{A}}^ 1_ k-\{0\},\) one takes \(K=(f_{\phi})_*{\mathcal L}_{\chi}(g_{\phi})[m]\), where this time \(\chi: {\mathbb{F}}^{\times}_ q\to {\mathbb{C}}\) is a multiplicative character, and thus obtains a majoration for trigonometric sums \(S=\sum_{x}\psi (a\cdot f_{\phi}(x))\chi (g_{\phi}(x))\) (with \(a\in {\mathbb{Z}}^ r\), sum taken over \(X({\mathbb{F}}_ q))\), namely \(| S| \leq const. (\sqrt{q})^ m\), with a constant determined by \(X_{{\mathbb{C}}}\), provided that a certain polynomial expression in a is not divisible by p. The paper also considers an analogous Fourier theory for \({\mathcal D}\)-modules instead of \(\ell\)-adic sheaves.

Reviewer: J.H.de Boer

MSC:

14F30	\(p\)-adic cohomology, crystalline cohomology
14G20	Local ground fields in algebraic geometry
43A32	Other transforms and operators of Fourier type

Keywords:

\(\ell \)-adic Fourier transformation

References:

[1]	A. A. Beilinson, I. N. Bernstein etP. Deligne, Faisceaux pervers, Conférence de Luminy, juillet 1981, Analyse et Topologie sur les expaces singuliers, I,Astérisque 100 (1982).
[2]	I. N. Bernstein, Modules over a ring of differential operators. Study of the fundamental solutions of equations with constant coefficients,Funct. Anal. 5 (1971), 1–16. · Zbl 0246.17008 · doi:10.1007/BF01075841
[3]	I. N. Bernstein, Lectures onD-Modules, Conférence de Luminy, juillet 1983, “ Systèmes différentiels et singularités {”, preprint.}
[4]	J. E. Björk,Rings of Differential Operators, North-Holland (1979).
[5]	J. L. Brylinski,Transformations canoniques, dualité projective, théorie de Lefschetz, transformations de Fourier et sommes trigonométriques, prepint de l’Ecole polytechnique (1983).
[6]	J. L. Brylinski, (Co)-homologie d’intersection et faisceaux pervers, Séminaire Bourbaki 1981–1982, no 585,Astérisque 92–93 (1982), p. 129–157.
[7]	J. L. Brylinski, A. S. Dubson etM. Kashiwara, Formule de l’indice pour les modules holonomes et obstruction d’Euler locale,C.R.A.S. 293 (30 novembre 1981), 573–576. · Zbl 0492.58021
[8]	P. Deligne, La conjecture de Weil II,Publ. Math. IHES 52 (1980), 313–428.
[9]	P. Deligne, Equations différentielles à points singuliers réguliers,Lecture Notes in Math.163, Springer Verlag (1970).
[10]	H. Hironaka, Resolution of singularities of an algebraic variety over a field of characteristic zéro, I et II,Annals of Math. 79 (1964), 109–326. · Zbl 0122.38603 · doi:10.2307/1970486
[11]	N. M. Katz, Sommes exponentielles, Cours à Orsay, automne 1979,Astérisque 79 (1980).
[12]	G. Laumon, Semi-continuité du conducteur de Swan (d’après Deligne), dans Caractéristique d’Euler-Poincaré, Séminaire E.N.S. 1978–1979,Astérisque 82–83 (1981), 173–219.
[13]	G. Laumon, Comparaison de caractéristiques d’Euler-Poincaré en cohomologie -adique,C.R.A.S. 292 (19 janvier 1981), 209–212. · Zbl 0468.14005
[14]	G. Laumon, Majoration de sommes exponentielles attachées aux hypersurfaces diagonales,Ann. scient. Ec. Norm. Sup., 4e série,16 (1983), 1–58.
[15]	M. Raynaud, Caractéristique d’Euler-Poincaré d’un faisceau et cohomologie des variétés abéliennes,Séminaire Bourbaki 1964–1965, exposé no 286, W. A. Benjamin (1966).
[16]	J.-P. Serre, Cohomologie galoisienne,Lecture Notes in Math. 5, Springer Verlag (1964). · Zbl 0143.05901
[17]	J.-P. Serre, Majoration de sommes exponentielles, Journées Arithmétiques de Caen,Astérisque 41–42 (1977), 111–126.
[18]	T. Ekedahl, On the adic formalism, to appear. · Zbl 0821.14010
[19]	E. L. Ince,Ordinary Differential Equations, Dover (1956). · Zbl 0063.02971
[20]	Z. Mebkhout, The Riemann-Hilbert problem in higher dimension,Proc. Conf. Generalized Functions Appl. in Math. Phys. (Moscow, Nov. 1980), Steklov Inst. 1981, 334–341.