• 検索結果がありません。

CONFIGURATION SPACES OF DIVISORS AND HOLOMORPHIC MAPS(Recent development of algebraic topology)

N/A
N/A
Protected

Academic year: 2021

シェア "CONFIGURATION SPACES OF DIVISORS AND HOLOMORPHIC MAPS(Recent development of algebraic topology)"

Copied!
5
0
0

読み込み中.... (全文を見る)

全文

(1)

CONFIGURATION SPACES OF

DIVISORS AND HOLOMORPHIC MAPS

MARTIN GUEST

ANDRZEJ KOZLOWSKI KOHHEIYAMAGUCHI The University of Rochester

ToyamaInternational University The University ofElectro-Communications

\S 1.

Introduction.

Spaces of holomorphic maps betweencomplex manifoldshave played

a

major role In

such diverse branches of mathematics

as

analysis, differential geometry, topology and

mathematicalphysics.

If$X\subset \mathbb{C}P^{n}$ is

a

projectivevariety

we

denote by $Ho1_{d}^{*}(S^{2},X)(Ho1_{d}(S^{2},X))$the space

of allbased(non-based)holomorphic maps$S^{2}arrow X$of degree$d$, where$S^{2}$ Isthe Riemann sphere,$S^{2}=\mathbb{C}\cup\{\infty\}$

.

For simplicity

we

shall

as

sume

thatthe degree$d$is

a

non-negative

integer. The corresponding space ofbased (non-based) continuous maps ofdegree $d$ is

denoted by $Map_{d}^{*}(S^{2},X)(Map_{d}(S^{2},X))$

.

In [S] Segal studied the inclusion map

$I_{d}$ : $Ho1_{d}^{*}(S^{2},\mathbb{C}P^{n})arrow Map_{d}^{*}(S^{2},\mathbb{C}P^{n})$

andshowedthat this inclusionmapis

a

homotopyequivalenoeuptodimension$d(2n-1)$

.

For any projective variety $X$ It is natural to consider the following

Problem. When is th$e$map$I_{d}$ : $Ho1_{d}^{*}(S^{2},X)arrow Map_{d}^{*}(S^{2}, X)$

a

homot

$opy$equivalence

up to

some

dimension $n(d),$ $such$ that$n(d)arrow\infty$

as

$darrow\infty$?

Segal’s results for $X=\mathbb{C}P^{n}$ have been generalized to the

case

when $X$ Is

a

Grass-manian,

or

more

generally,

a

flag manifold (see [G], [Gu), [K], $[M^{2}]$). In this note

we

consider the

case

where $X$ is the complement of

a

union of linear subspaces in $CP^{n}$:

$X= \mathbb{C}P^{n}\backslash \bigcup_{\alpha\in\Lambda}H_{\alpha}$, where $\{H_{\alpha} :\alpha\in\Lambda\}$ is

a

family of linearsubspaces of$\mathbb{C}P^{n}$

.

Our

purpose is to

announce

the main results of [GKY], which extend Segal’s results [S] to

the

case

$X=X_{n}= \mathbb{C}P^{n}\backslash \bigcup_{0\leq i<j\leq n}H_{i,j}$, where$H_{i,j}=\{[z_{0}$ : $z_{1}$ :... : $z_{n}]\in \mathbb{C}P^{n}$ : $z_{i}=$

$z_{j}=0\}$

.

(2)

Theorem 1. The inclusion $maps$

$I_{d}$ : $Ho1_{d}^{*}(S^{2}, X_{n})arrow Map_{d}^{*}(S^{2},X_{n})$ and

$J_{d}$ : $Ho1_{d}(S^{2},X_{n})arrow Map_{d}(S^{2},X_{n})$

are

homologyequivalences up to dimension$d$

.

Theorem 2. If$2d>n$ thetwo mapsabove

are

homotopy$equ\dot{!}valences$up todimension

$d$

.

Here

we

$caU$

an

inclusionmap$Xarrow Y$

a

homotopyequivalence (homoloqy equivalence)

up to dimension $m$ if$\pi_{j}(Y, X)=0$ when$j\leq m$ (if $H_{j}(Y,X)=0$when$j\leq m$).

Remark.

(1) For$n=1$ the above raeul$ts$

were

obtained in [S].

(2) We expect that similar$m$ethods

can

be used to obtain analogousresults when

$X= \mathbb{C}P^{n}\backslash \bigcup_{I}P(I)$, where$P(I)=\{[z_{0} : \cdots : z_{n}]\in \mathbb{C}P^{n} : p_{j}=0ifj\in I\}$, and

the union is

over a

collection ofsubsets$I$ of$\{0,1,2, \ldots , n\}$

.

\S 2.

Configuration Spaces of Divisors.

Definition 2.1. For

a

connectedpair of CW-complexes (X, Y), let $Sp^{d}(X, Y)$ denote

the d-fold symmetric product of$X/Y$

.

Adding

a

base point gives rise to

a

natural

inclusion $Sp^{d}(X, Y)arrow Sp^{d+1}(X,Y)$ and

we

put $Sp^{\infty}(X, Y)=\bigcup_{d\geq 1}Sp^{d}(X, Y)$

.

We

define

a

space $Q_{d}^{(n)}(X,Y)$ by

$Q_{d}^{(n)}(X,Y)=$

{

$(\xi 0,$$\ldots,\xi_{n})\in(Sp^{d}(X,Y))^{n+1}$ : $\xi:\cap\xi_{j}=\emptyset$ if$i\neq j$

}.

If$Y=\emptyset$,

we

write$Sp^{d}(X)=Sp^{d}(X, \emptyset)$ and$Q_{d}^{(n)}(X)=Q_{d}^{(n)}(X, \emptyset)$

.

If$M$is

a

connectedopen manifold,adding $(n+1)$ distinctpoints (ffominfinity”$(c$

.

$f$

.

[Mc]) gives

a

natural stabilizationmap$i_{d}$ : $Q_{d}^{(n)}(M)arrow Q_{d+1}^{(n)}(M)$ and

we

define$\hat{Q}^{(n)}(M)$

to be the “identity component” of $\lim_{darrow\infty}Q_{d}^{(n)}(M)$

.

Let $F(X, m)$ be the configuration

spaceofm-tuplesof distinct pointsin $X$

.

In particular, $Q_{1}^{(n)}(X)=F(X, n+1)$, and it

is well-known that $\pi_{1}(F(\mathbb{C}, m))=I(m)$, where $I(m)$ denotes the group of pure braids

of$m$ strings. Then

we

have

Proposition 2.2.

(1) $Ho1_{d}^{*}(S^{2},X_{n})=Q_{d}^{(n)}(\mathbb{C})$

.

(2) $\pi_{1}(Ho1_{d}^{*}(S^{2},X_{n}))=\{\begin{array}{l}I(n+1)jfd=1\mathbb{Z}^{n(n+1)/2}ifd\geq 2\end{array}$

(3)

Theorem3.1. ([GKY],[Ko]). Thestabilizationmap : $Ho1_{d}^{*}(S^{2}, X_{n})arrow Ho1_{d+1}^{*}(S^{2},X_{n})$

is ahomology$eq$uivalence up to dimension$d$

.

Usingthe McDuff-Segal transfer ([Mc],[S])

we

obtain

Proposition 3.2. For any

comm

utative ring $R$, the induced homomorphism

$i_{d*}$ : $H_{*}(Ho1_{d}^{*}(S^{2}, X_{n}),$ $R$

)

$arrow H_{*}(Ho1_{d+1}^{*}(S^{2},X_{n}),$$R$

)

is

a

spli$t$ monomorphism. More

precisely, there is

a

familyof graded R-modules $\{R_{m}:’m\geq 0\}$ such that

(a) $H_{*}(Ho1_{d}^{*}(S^{2}, X_{n}),$ $R$

)

$= \bigoplus_{0\leq m\leq d}R_{m}$

.

(b) The aboveisomorphism is compatible with the splitting monomorphism.

These results lead

us

to expect

Conjecture 3.3. Thereis

a

stable splitting

$Ho1_{d}^{*}(S^{2},X_{n})\simeq\vee D_{j}(n)S1\leq j\leq d$

such that

$D_{j}(n)\simeq Ho1_{j}^{*}(S^{2},X_{n})/Ho1_{j-1}^{*}(S^{2},X_{n})s$

Remark 3.4. $([C^{2}M^{2}])$ This is true for $n=1$

.

\S 4.

The Scanning Map.

Definition 4.1. Let $\epsilon>0$ beanypositiverealnumber, and let $D_{z}(\epsilon)$ deno$te$ theopen

disk ofradius$\epsilon$ with centreat $z\in \mathbb{C}$

.

Define themap $S_{d}$ : $Q_{d}^{(n)}(\mathbb{C})x\mathbb{C}arrow Q^{(n)}(S^{2}, \infty)$

by

$((\xi 0, \ldots,\xi_{n}), z)arrow\rangle(\xi 0\cap D_{z}(\epsilon), \ldots,\xi_{n}\cap D_{z}(\epsilon))\in Q^{(n)}(\overline{D}_{z}(\epsilon),\partial\overline{D}_{z}(\epsilon))$

$\simeq Q^{(n)}(S^{2}, \infty)$

.

Since $\lim S_{d}(\Xi, z)=(\emptyset, \emptyset, \ldots, \emptyset)$for$any\Xi\in Q_{d}^{(n)}(\mathbb{C})$,

we

defin$eS_{d}(\Xi, \infty)=(\emptyset, \emptyset, \ldots, \emptyset)$

and $obtainarrow\infty$

a

map

$S_{d}$: $Q_{d}^{(n)}(\mathbb{C})\cross S^{2}arrow Q^{(n)}(S^{2},\infty)$

.

Taking the adjoint

we

obtain

a

$map$

(4)

Itshomotopy class is independent of the choice of$\epsilon$. We call$S_{d}$ the

scannin.

$q$ map.

It

can

be shown that $Q^{(n)}(S^{2}, \infty)\simeq n+1\vee \mathbb{C}P^{\infty}$

.

It is also easy to

see

that there Is

a

homotopy equivalence $\alpha_{d}$ :

$\Omega_{d}^{2}(\vee \mathbb{C}P^{n})n+1\simeq\Omega_{d+1}^{2}(\vee \mathbb{C}P^{n})n+1$ such that the following

diagram is commutative up to homotopy

$Q_{d}^{(n)}(\mathbb{C})arrow^{s_{d}}$ $\Omega_{d}^{2}(\vee \mathbb{C}P^{\infty})n+1$

$Q_{d+1}^{(n)}(\mathbb{C})\downarrow i_{d}arrow^{s_{d+1}}\Omega_{d+1}^{2}(\vee \mathbb{C}P^{\infty})\simeq\downarrow dn+1^{\alpha}$

Considerthe mapping telescope of the maps

$Q_{1}^{(n)}(\mathbb{C})arrow i_{1}Q_{2}^{(n)}(\mathbb{C})arrow:_{2}Q_{3}^{(n)}(\mathbb{C})arrow:_{3}Q_{4}^{(n)}(\mathbb{C})arrow\ldots$

It is easy to

see

that this mapping telescopeis homotopy equivalent to $\hat{Q}^{(n)}$

.

Hence

we

obtain

a

stabilized scanningmap

$\hat{S}$: $\hat{Q}^{(n)}arrow\Omega_{o(\vee \mathbb{C}P^{\infty})}^{2^{n+1}}$

.

By arguingexactly

as

in [S],

we

obtain

Proposition 4.3. Thescanningmap $\hat{S}$

is

a

homotopyequivalence.

Sketchprvofs

of

Theorems 1 and 2. Let $G=(\mathbb{C}^{*})^{n}$ and define

a

G-action

on

$X_{n}$ by $((t_{1}, \ldots,t_{n}), b:\cdots : p_{n}])rightarrow\beta n:t_{1}p_{1}$ : $\cdots$ : $t_{n}p_{n}$].

Then there is

a

fibre sequence

$\mathcal{I}^{m}arrow X_{n}arrow qn+1\vee \mathbb{C}P^{\infty}$

.

(This follows Rom the fact that $EG\cross GX_{n}\simeq n+1\vee \mathbb{C}P^{\infty}$).

There is

a

homotopy

commu-tative diagram:

$Ho1_{d}^{*}(S^{2}, X_{n})arrow^{I_{d}}Map_{d}^{*}(S^{2},X_{n})=\Omega_{d}^{2}X_{n}$

$\simeq\downarrow$ $\simeq\downarrow\Omega^{2}q$

$Q_{d}^{(n)}(\mathbb{C})$ $arrow^{s_{d}}$ $\Omega_{d}^{2}(^{n+1}\mathbb{C}P^{\infty})$

It follows that $\lim_{darrow\infty}I_{d}$ is

a

homotopy equivalence. Hence Theorem 1 follows from the

stabilization theorem.

Finally,

an

argument analogous to the

one

given by Segal in [S] shows that the space

$Q_{d}^{(n)}(\mathbb{C})$ is nilpotent uptodimension$d$if$2d>n$

.

Theorem2 follows fiiom the Whitehead

(5)

REFERENCES

$[C^{2}M^{2}]$ Cohen,F. R.,Cohen,R. L.,Mann,B. M.andMilgram, R. J.,Thetopologyofrationalfunctions

and divisors ofsurfaces, Acta Math. 166 (1991), 161-221.

[CLM] FRCohen, Lada, T. Y., May, J. P., The homology ofiterated loop spaces, vol. 533, Springer LectureNotes In Math., 1976.

[C-S] R L Cohen and DHShimamoto, Rational functions, $\iota_{abe}u_{ed}$ configurations and Hilbertschemes,

J. London Math. Soc.43 (1991), 505-528.

[E] Epshtein,S. I., thndamentalgroups ofspaces ofcoprimepolynomials, FunctionalAnalysis andIts

Applications 7 (1973), 82-83.

[Gj Gravesen, J., On the topology ofspaces ofholomorphic maps,ActaMath. 162 (1989), 247-286.

[Gu] Guest, M. A., Topology ofthe space of absolute minima of the eneryfunctional, Amer. Jour. Math. 106 (1984), 21-42.

[$c\kappa\eta$ Guest, M. A., Kozlowski,A.and Yamaguchi, K., The topologyofspacesofcoptmepolynomials, Preprint.

[HR] Hilton,P.and Roitberg, J.,On the Zeemancomparisontheoremforthe homologyofquasi-nilpotent fibmtions, QuarterlyJ. Math. 27 (1976), 433-444.

[K] Kirwan, F., On spaces ofmapsfrom Riemann sphares to Grassmanians and applications to the cohomology ofmoduli ofvector bundles, Ark. M. 24 (1986), 221-275.

[Ko] Kozlowski, A., Stabilization ofhomology groups ofspaces ofmutually disjoint divisors, Inthese Proceedings.

$[M^{2}]$ Mann, B. M. and Milgram, R. J., Some spacesofholomorphicmapsto complexGmssman

man-ifolds,J. Differential Geometry33 (1991), 301-324.

[Mc] McDuff,D., Configurationspacesofpositiveand negative particles, Topology 14 (1975), 91-107.

参照

関連したドキュメント

Beer introduced the problem of the global coincidence on C(X, Y ) for metric spaces, and proved that if the metric space Y contains a non trivial arc, than the above two

The main problem upon which most of the geometric topology is based is that of classifying and comparing the various supplementary structures that can be imposed on a

We know that the function u ˜ i is p(·)-quasicontinuous; notice here that [21], Theorem 2, improves [15], Theorem 4.6 by showing that our standard assumptions are sufficient for

We shall refer to Y (respectively, D; D; D) as the compactification (respec- tively, divisor at infinity; divisor of cusps; divisor of marked points) of X. Proposition 1.1 below)

In Section 3 we use this set up to prove Theorem 3.1, which characterizes the boundaries of graphs of harmonic functions using the moment conditions arising from conservation laws..

Classical definitions of locally complete intersection (l.c.i.) homomor- phisms of commutative rings are limited to maps that are essentially of finite type, or flat.. The

Now it follows from a similar argument to the argument used in the proof of [7], Theorem 4.1, that the images in Π X (r+1) of the geometrically pro-l log fundamental groups of

Yin, “Global existence and blow-up phenomena for an integrable two-component Camassa-Holm shallow water system,” Journal of Differential Equations, vol.. Yin, “Global weak