grupo de fisica matematica
TRANSCRIPT
Grupo de Fisica Matematica Universidade de Lisboa
00 •love"'"� 0I -~ \..l
1-:1
The Renormalization of Self Intersection Local Times
I: The Chaos Expansion
~- Margarida de Faria, Custodia Drumond and Ludwig Streit
IFM 1/98
. -'. !"\ - ~1:lS- 'J 07 90 • FAX 35LLJ-'15..A288n~28~=.J....J.S.1
The Renormalization of Self Intersection Local Times
I: The Chaos Expansion
Margarida de Faria* Custodia Drumond*� Ludwig Streit*+�
*CCM, Universidade da Madeira, P 9000 Funchal� http://www.uma.pt/�
+BiBoS, Univ. Bielefeld, D 33615 Bielefeld�
June 17, 1998
1. Introduction
The intersections of Brownian motion paths have been investigated since the Forties [20]. One can consider intersections of sample paths with themselves or e.g. with other, independent Brownian motions [36], one can study simple [5] or n-fold intersections [6] and one can ask all of these questions for linear, planar, spatial or - in general- d-dimensional Brownian motion: evidently self-intersections become increasingly scarce as the dimension d increases.
Intersection local times of Brownian motion were studied by many authors, see e.g. [1], [3]-[11]' [15], [18]-[38]. A more systematic review than we can give here of the subject will be found e.g. in the recent reference [15].
An informal but rather suggestive definition of self intersection local time of Brownian motion B is in terms of an integral over Dirac's - or Donsker's - 8function
; L = Jd2t 8(B(t2) - B(t l )),
intended to sum up the contributions from each pair of "times" t l , t2 for which the Brownian motion B is at the same point.
- -
In Edwards' modeling of polymer molecules by Brownian motion paths, L is used to model the" excluded volume" effect: different parts of the molecule should not be located at the same point in space. As another application, Symanzik introduced L as a tool for constructive quantum field theory in [31].
A rigorous definition, such as e.g. through a sequence of Gaussians approximating the 8-function or in terms of generalized Brownian functionals [10] [30] [33], will lead to increasingly singular objects and will necessitate various "renormalizations" as the dimension d increases. For d > 1 the expectation will diverge in the limit and must be subtracted [18] [32], clearly L will then no more be positive. For d > 3,5,7, ... further subtractions have been proposed [33] that will make L into a well defined generalized function of Brownian motion.
For d = 3 another renormalization has been constructed by Westwater to make the Gibbs factor e-g
•L of the polymer model well-defined [35], for yet another,
recent approach see [1]. Yor, in [38], first suppresses the short time accumulation of self-intersections
by the regularization
8(B(t- 2) - -B(t1 )) ~.8(B(t2) - -B(t1) + c)
and shows, again for d = 3, that a multiplicative renormalization
gives rise to another, independent Brownian motion as the weak limit of regularized and subtracted approximations to L.
In the present note we study similar limits, for arbitrary d ~ 3, using a Gaussian regularization of the 8-function for which the chaos expansion of the corresponding regularized L e is available [10]. For a suitably subtracted and renormalized local time, each term in this expansion converges in law to a Brownian motion.
We prepare and state these results in the following section 2, in section 3 we give their proofs. In a companion paper [3] we extend these results to the corresponding series.
2. Definitions and Main Results
2.1. White Noise Analysis and Local Times
We reproduce here some White Noise Analysis concepts as introduced in [10], referring to [14] for a systematic presentation.
2
Brownian motions Bi , i = 1, ... , d, have version in terms of white noise Wi via
Hence we consider independent d-tuples of Gaussian white noise w = (WI, ... ,Wd) and correspondingly, d-tuples of test functions f = (il, ... , Id) E S(R, Rd), and introduce the following notation:
d d
r;,= (nI, ... ,nd), n = L ni, r;" = IT ni! I I
d
< f, f > = L Jdt H(t) i=1
and similarly for <: w®n' :, Fn' > where for d-tuples of white noise the Wick product: ... : [14] generalizes to
- d w0ni. w0n '= JO,. .• . \CI. t ..
i=1
The vector valued white noise w has the characteristic function
C(f) = E(ei<w,f» = r dJ.L [w] ei<w,f> = e-~<f,f>, (2.1)j S·(R,Rd)
d
where < w, f > = I: < Wi, Ii> and Ii E S(R, R). i=1
The Hilbert space (L2
) = L 2 (dJ.L)
is canonically isomorphic to the d-fold tensor product of Fock spaces of symmetric square integrable functions:
(£2) ~(; SymL2(Rk, k!dkt))0d = J, (2.2)k=O
3
for the general element of (L2 ) this implies the chaos expansion
cp(w) = I:00
<: w®r; :, Fr; > (2.3)
with kernel functions F in J. It is desirable to introduce regularizations for the intersection local time, with
a view towards the construction of well-defined, "renormalized" intersection local times in higher dimensions where the latter do not exist without subtractions. A computationally simple regularization is, for c > 0,
with 1B 2be (B (t2) - B (t1)) =(21rc)-d/2 e_ (t );:(t,)I2 (2.4)
It has the following chaos expansion, which we quote here only for d ~ 3 :
Theorem 2.1. [10J For any c > 0, Le - E (Le) has kernel functions FE J given by
(2.5)
8(u)8(t - v) . ((v - u + c)-X: + (t + c)-X: - (v + c)-X: - (t - u + c)-X:)
ifall Ili are even, and zero otherwise, with V(Sl, ... , sn) =max(sl, ··'l sn), U(Sl, ... , sn) = min(sll '.'l sn), and x = (n + d) /2 - 2. 8 is the Heaviside function.
Each kernel function is thus the sum of four terms. The first one gives rise to
Definition 2.2.
Mt(d, rt, c) = r crs(v - U + c)-X: : w®n' (s) : (2.6)� J[O,tl n�
d� _ r crs(v - U+ c)-X: Q9 :wi(si) ...Wi(S~) :� J~~n i=l�
4�
1
d ni t Si d
- L L 1ds~ 1'" ~-lS(V - U + c)-X 0 :Wi(S;) ...Wi(S~J : i=l m=l 0 0 t=l
- t ni it dBi(T) iT ~-lS(T - U + c)-X : w®~(s) : i=l 0 0
d t
- L 1dBk(v)mk(V) (2.7) k=l 0
- Ld
Mk,t (2.8) k=l
The others give
Nt(d,n:,c) _ r dns((t+c)-X_(v+c)-X_(t-u+c)-X) :w®n(s): Jro,tjn
All of the above processes are continuous.
Definition 2.3. We denote the nth order contribution to the regularized local time Lf; by
Remark 1. Our key observation is that, as c goes to zero, 1\11 is more singular than N, and that it is a Brownian martingale.
2.2. The main theorems
Theorem 2.4. For d 2: 3, the renormalized M converge in distribution to independent Brownian motions {3i:
with 2 ---. { n(n - 1) if d = 3
kn = n! n!(d-4)! '[ d 3 (2.9) (n+d-5)! 1 >
5
while for d = 3
II(v - U + E:)-XII~2([o,tln) = n(n - l)t IInE:1 (1 + 0(1))
•To show convergence of these martingales by Theorem VIII.3.11 of [16] we
must show convergence of characteristics. Since the processes M are continuous this reduces to showing convergence in probability of (rI\lfi , rl'Vfk)t as E: -'> +0. This will be taken care of by proposition 3.6 which we prepare now:
(rl'Vfi,rlvh)t = r20ik it dv(mi(v))2
and we need to estimate
Let us note that
2E ((rMi, rMk)t) = E (r MiMk) = Oikr2 : k n! II(v - U + E:)-XII~2([o,t)n)
E ((rMi, rMk)t) = Oik nk k;' (t + 0(1)) n
Next we intend to show that the rhs gives indeed that the rest of (rMi, rl'Vfk)t goes to zero. The kernel of the highest order is
C(i) ( I. . ')( )-X( I )-X') -ltd 8(---+ ---+ Sl'Sl" ... ,Sn-l,Sn_l - T T-VVV T-U+E: T-U +E:2n -2 {j 0i
(3.1)
V(/) and U (I) are the largest and the smallest of the S~/), and (7!i) k == Oik.
(For n = 2: U(/) = v(1) = s(1)).
Remark 2. The integral (3.1) can be calculated in closed form (using e.g. {12} nos. 2.15 and 2.263.4), but one gets a useful approximation by introducing the following auxiliary functions:
H2n- 2 (Sl, ... , Sn-l; s;, ...,S~_l; E:; d) = (v V Vi - U V u' + E:)-(n+d)/2+3(v V Vi - U /\ u' + E:)-(n+d)/2+2
where v =max(si), U =min(si) and v' =max(s~), u' =min(sD. These functions majorize the kernel functions
7
Lemma 3.3. For n > 2, and x > 1
O G (i) (' I )< 2n-26i 81,81;, ; 8n -1, 8 n -1 (3.2)
< ~l H2n- 2(81 , , 8n-1; 8~, ... , 8~_1; e; d)x-
Proof:
o < ltdT8(T-vVVI)(T-u+e)-X(T-ul+e)-X
< it dT(T - U V u' + e)-X(v V v' - u A u' + e)-X vvv' 1
< --(v V v' - u V u' + e)-x+l(v V v' - u A u' + e)-Xx-l
This estimate suffices to show
--+ --+Lemma 3.4. For 2 n - 28i =1= 0
r2G(~ __ -+ 0 in £2(R2n- 2)2 n -2 8 i
as e goes to +0.
Proof: Consider first n 2: 3. By the above estimate it is sufficient to show that
lim r4 II H2n- 211 ~2([0 t]2n-2) = o. e:->+O '
IIH2n-211~2(R2n-2) - J~-1 8 J~-1 8' (V V v' - u V u' + e)2-2x(v V V' - u A u' + e)-2x
t
du lvl
- Cn l dv l v
dv' l v
du'
(v _u)n-3(v'_u,)n-3
(V V v' - u V u' + e)n+d-6(v V v' - u A u' + e)n+d-4
< Cne8-2d I t
/e: dy lY
dy' lY
dx ly' dx'
1 (y - x V x' + 1)d-3(y - x A x' + 1)d-1
8
and then decompose the x-integration into the following two domains
X' < x and x < x'
In the first case
8 I e -
2d it/€: dy i Y
dy' i yl
dx' 1~ dx (y _ x + 1)d-3~y _ x' + l)d-l
8 /1- e -2d it/€: dy i
Y
dy' i yl
dx' (y _ X + l)d-l 1~ dx (y _ x ~ 1)d-3
We first estimate the last integral
y - x' if d = 3 In (y - x' + 1) if d = 41~ -(Y---X-~-l)-d--3 dx ~ { d~4 if d> 4
and one finds 0 (1) if d = 3
I = { 0(c7-2d) if d> 3
Hence, in both cases, r 4 I vanishes as e -- O. The second case is
t c l yl� 8 2d� / 1 1I = e - I dy lY dx ( )d-l l Y
dy' dx' ( I )d-3 (3.3) o 0 y-x+l x x y-x+1
Estimating the last integrand by 1 we find
Y y' ( )2
1d I 1 dx' 1 < y - x x y x (y - x' + 1)d-3 - 2
Substitution of these estimates into the integrals over x and y gives for d 2: 3
It/€: lY ( )2 if d ={ 0(1) 3
I :s const.e8-
2d dy dx ( y - X )d-l = O(c-lIne) if d = 4 o 0 Y - x + 1 O(c7 - 2d ) if d > 4
9�
so in that r 4 I vanishes as c -+ O. For n = 2 it is sufficient to use the estimate
o < G2 = it d78 (7 - U V u') (7 - u+ c)1-d/2(7 - U' + c)1-d/2
< it d7(7 - u V u' + c)1-d/2(u V U' - u /\ u' + c)1-d/2 = H 2(u, u') uVu'
and to verify that
•With this lemma we have established that the highest order term of the (renormalized) quadratic variation goes to zero in quadratic mean for any t > O. The kernels Gk of the other terms are obtained by integrating over pairs of s, s' such as e.g. m
(i) .'• . ,it
Sym 0 dsG2(n_6.) (s, S, S2, S2""" Sn-l, sn_l)'
these new functions are in fact also bounded by an expression like (3.1), and hence by (3.2), so that for all c > 0,
With this goal we show
Lemma 3.5. Let n ;::: 2 and m > 1 , and let Fn,m be a function symmetric in the variables S and in the variables s', with
o < Fn,m(Sl, ... ,Sn-l,Sn,S~,.",S~_l'S~) (3.4)
< cit d78 (7- max (s, S')) (7- min (Si) + c)-m(7- min (s~) + c)-mo I~n I~n I~n
Then 3em < 00 such that
o < it dsFn,m(Sl,' .. ,Sn-l, S, s~, . .. ,S~_l' s) (3.5)
< em t d78 (7- n:tax (s, S')) (7- ~in (Si) + c)-m+l/2(7_ min (s~) + c)-m+l/2Jo I<n I<n I<n
10
l !
Proof: Under the assumption of the lemma
t
f dsFn,m (SI' ... , Sn-l, S, S~, ... , S~_I' S) :S o
t t
cfdsJ'dr8(r- max (Si'S~,S)) ( r- min (Si'S) + c) -m (r- min (S~, s) + c)-m O~t~n-l O~t~n-l O~t~n-l
o 0
Using u = min Si , u' = min S~
O~i~n-l O~i~n-l t
V = max Si Vi = max s' 09~n-l ' O~i~n-l t
(For n = 2: U(I) = V(I) = sf)). Assuming without loss of generality that u < u' we can decompose the s-integration of our estimate as follows
t t
C f ds f dr8 (r - max (v, Vi, s)) (r - min (u, s) + c) -m (r - min (u' , s) + c)-m
o 0
·8 (r - max (v, Vi, s)) (r - min (u, s) + c)-m (r - min (u' , s) + c)-m
- c j dr (J ds(r-s+€)-2m+ Jds(r-u+€)-m(r-s+€)-m�
vVv' 0 u�
+ "]' ds (r - u + €)-m (r - u' + €)-m + Jrds (r - u + €)-m (r - v! + e)-m)
~ vv~
We now show that each of the four terms obeys the postulated estimate (3.5)
t u t
Jdr Jds (r - S+ c) -2m :S 1 Jdr (r - u + c) -2m+l 2m-l
vvv' 0 vVv'
11
t
:::; 2m1_ 1 j d78 (7 - Vv V') (7 - u + C)-m+~ (7 - U' + C)-m+~
o
since u < u' and m > 1/2. The second term
t u'
j d7 j dS(7-U+c)-m(7-S+C)-m
vvv' u�
t�
< 1 jd7(7-U+C)-m(7-U'+C)-m+lm-1
vVv'� t�
< 1 j d78 (7 - VV V') (7 - u + c)-m+~ (7 - U' + c)-m+~ . m-1
o
using again u < u'. The third term
t vvv'
j d7 j ds (7 - U+ c) -m (7 - U' + c)-m
vvv' u'�
t�
- j d7 (7 - U + c) -m (7 - U' + c) -m (V V V' - u')
vvv'� t�
< j d7 (7 - U + c) -m (7 - U' + c) -m (7 - u')
vVv'� t�
< j d78 (7 - V V V') (7 - u + c)-m+~ (7 - U' + c)-m+~ .
o
Finally
t
j d7 j tds (7 - U + c)-m (7 - U' + c)-m
vvv' vvv'
12
tJdr (r - u + c;)-m (r - u' + c;)-m (r - v V v')
vVv'� t�
< Jdr (r - u + c;) -m (r - u' + c;) -m (r - u')
vVv'� t�
< Jdr8 (r - v V v') (r - u + c;)-m+~ (r - u' + c;)-m+~
• o
Combining this Lemma with the previous one we conclude that for all kernel functions Gk with k 2: 2 arguments
lim r411SymGkll2 ~ lim r411Gkll2 ~ lim r411Hkl12 = a e-++O e-++O e-++O
i.e. we have shown Proposition 3.6.
Proof of Theorem 2.4: The above limit is clearly (up to constants) the quadratic variation of a Brownian motion. Theorem 2.4 is then a consequence of e.g. Theorem VIII.3. 11 in [16], which in the present case of continuous martingales requires convergence in probability of quadratic variations for a dense set of t.
To control the remaining terms Nt(d, Tt, c;;) in the chaos expansion we observe that
IINt(d, Tt,c;)II~L2) - Tt! 11(t + c;)-X - (v + c;)-X - (t - u + c;)-XII~2([o,tJn)
< Tt! (II (t + c;)-XII~2 + \I (v + c;)-XII~2 + II (t - u + c;)-XII~2)
The first of these three norms is equal to tn(t + C;)-2x, i.e. 0(1). The second one IS
it / e 1t vn-l xn i4 dr ~s(v + C;)-2x = n dv = nc - dx-:---""'7""""'"~-:-
J[O,tl n o (v + c)n+d-4 0 (x + 1)n+d-4
0 (1) for d = 3 - O(lnc) for d = 4
{ 0(c4- d ) for d > 4
13
which are suppressed by the renormalization
1 r2(e) = {llne l - for d = 3.
ed- 3 for d> 3
A similar estimate holds for the third term of N, so that we have shown
Lemma 3.7.
ms - lim r(e)Nt(d, Tt,e) = O. e:--+O
In fact the convergence is uniform in any finite t-interval. Next we show
Lemma 3.8. The processes {r(e)N.(d, Tt,e;) : e > O}, {r(e)M.(d, Tt,e;) : e > O}� and their linear combinations are tight.�
Proof: A criterion for tightness of M (following [17], p.64) is�
sup E IrMt - rMs\O< S CT (t - s)1+13 , (3.6) e:>0
\IT> 0 and 0 ::; s < t ::; T and for some positive constants a, (3 and CT' As a first step we show
sup E IrMt - rMsl 2 S CT (t - s) (3.7)
e:>0
by direct calculation:
v2E IMt-M s l = Tt!n it dv l cr-1s (v - u + e)-2x
The second integral may be estimated as follows v� lv (v - u)n-2cr-Is (v - U + e)-2)( = (n - 1) du n+d-4
O o (v-u+e)
< (n - 1)c3-d lv/e: dx 1 d-2
lo (x + 1)
_� (n _ 1) { In (v + e) - In e for d = 3 O(e3-
d) for d> 3
14
Renormalization of this estimate by the factor r 2 makes it bOllllded on [0, T], and the integral over v gives the desired estimate (3.7), Le.
(3.8)
Note that the kernel functions of K are all dominated by those of M. Hence we have also, possibly with a larger constant CT, the estimate
(3.9)
By the hypercontractivity of the Ornstein-Uhlenbeck semigroup (see e.g. [14], p.235), one has for nth order white noise monomials r.p E (L 2 ), and any a > 2
For r.p = rK t - rK s ,and using the above estimate for the 2-norm, we get
E IrKt - rKsl Q
~ CT (t - st/2
as required to ensure tightness. The estimates for rN etc. are of the same kind.• Proof of Theorem 2.5: We need to consider
rK=rM+rN
knowing that, as c -+ +0, the rK are tight by Lemma 3.8, the rM converge in law, and the rNt go to zero in mean square. The latter two facts are sufficient, via the Cramer-Wold device (see e.g. [17] p.61), for finite dimensional convergence of rK; tightness then implies convergence in law.
Acknowledgements: This work has had partial support from PRAXIS XXI and FEDER. L. S. would like to express his appreciation for an inspiring discussion with L. Chen, for a very helpful remark of J. Potthoff, and for the splendid hospitality of the Grupo de Ffsica Matematica da Universidade de Lisboa, llllder the auspices of a "Marie Curie" fellowship (ERBFMBICT 971949).
References
[1]� Bass, R. F., Khoshnevisan, D.: Intersection Local Times and Tanaka Formulas. Ann. Inst. H. Poincare 29 (1993) 419-451.
[2]� Deck Th., Potthoff J., Vage G.: A Review of White Noise Analysis from a Probabilistic Standpoint. Acta Appl. Math. 48 (1997), 91-112
15
[3]� Drumond C., de Faria M., Streit L.: The Renormalization of Self Intersection Local Times II: The Renormalized Local Times. In preparation.
[4]� Dvoretzky, A., Erdos, P., Kakutani. S. Double points of paths of Brownian motion in the plane. Bull. Res. Council Israel Sect. F 3 (1954), 364-37l.
[5]� Dvoretzky, A., Erdos, P., Kakutani, S. Double points of paths of Brownian motion in n-space. Acta Sci. math. Szeged, 12 (1950) 75-8l.
[6]� Dvoretzky, A., Erdos, P., Kakutani, S., Taylor, S. J. Triple points of the Brownian motion in 3-space. Proc. Cambridge Pillios. Soc. 53 (1957), 856862.
[7]� Dynkin. E. B.: Polynomials of the occupation field and related random fields. J. Funct. Anal. 58 (1984), 20-52.
[8]� Dynkin. E. B.: Self-intersection gauge for random walks and for Brownian motion. Ann. Probab. 16 (1988), 1-57.
[9]� Dynkin. E. B.: Regularized self-intersection local times of planar Brownian motion. Ann. Probab. 16 (1988),58-73.
[10]� de Faria M., Hida T., Streit L., Watanabe H.: Intersection Local Times as Generalized White Noise Functionals , Acta Appl. Math. 46 (1997), 351-362.
[11]� Geman. D., Horowitz, J., Rosen, J.: A local time analysis of intersections of Brownian paths in the plane. Ann. Probab. 12 (1984), 86-107.
[12]� Gradshteyn I. S., Ryshik I. M.: "Table of Integrals, Series, and Products" Corrected and enlarged edition. Academic Press, 1980.
[13]� Hida T.: "Brownian Motion". Springer, Berlin, 1980.
[14]� Hida, T., Kuo, H.-H., Potthoff, J. and Streit, L.: "White Noise - An Infinite Dimensional Calculus". Kluwer-Academic, 1993.
[15]� Imkeller, P., Perez-Abreu, V., Vives, J.: Chaos expansions of double intersection local times of Brownian motion in R d and renormalization. Stoch. Proc. Appl. 56, (1995) 1-34.
[16]� Jacod J., Shiryaev A. N.: "Limit Theorems for Stochastic Processes" Springer, Berlin, 1987.
16
[17]� Karatzas 1., Shreve S. E.: "Brownian Motion and Stochastic Calculus" 2nd ed., Springer, Berlin, 1991.
[18]� Le Gall, J. F.: Sur Ie temps local d'intersection du mouvement brownien plan et la methode de renormalisation de Varadhan. Sem. de Prob. XIX, 1983/84. LNM 1123 (1985), 314-331. Springer, Berlin.
[19]� Le Gall, J. F.: Sur la saucisse de Wiener et les points multiples du mouvement brownien. Ann. Probab. 14 (1986), 1219-1244.
[20]� Levy, P.: Le mouvement brownien plan. Amer. J. Math. 62 (1940), 440-487.
[21]� Lyons, T. J. The critical dimension at which quasi-every Brownian motion is self-avoiding. Adv. in Appl. Probab. (Spec. Suppl. 1986), 87-99.
[22]� Nualart, D., Vives, J.: Smoothness of Brownian local times and related functionals. Potential Analysis 1 (1992), 257-263.
[23]� Nualart, D., Vives, J.: Chaos expansion and local times. Publications Mathematiques 36(2), 827-836.
[24]� Penrose, M. D.: On the existence of self-intersections for quasi-every Brownian path in space. Ann. Probab. 17 (1989), 482-502.
[25]� Perez-Abreu, V.: Chaos Expansions: A Review, CIMAT preprint, 1993.
[26]� Rosen, J.: A local time approach to the self-intersections of Brownian paths in space. Comm. Math. Phys. 88 (1983),327-338.
[27]� Rosen, J.: Tanaka's formula and renormalisation for intersections of planar Brownian motion. Ann. Probab. 14 (1986), 1425-1251.
[28]� Rosen,. J.: A .renormalized local time for multiple intersections of planar Bro.wnwn mo~wn. Sem. de Prob. XX, 1984/85. LNM 1204 (1986), 515-53l. Sprmger. Berlin.
[29]� Shieh, N. R.: White noise analysis and Tanaka formula for intersections of planar Brownian motion. Nagoya Math. J. 122 (1991), 1-17.
[30] Streit, L. and Westerkamp, W.: A generalization of the characterization theorem for generalized functionals of white noise. in "Dynamics of Complex and Irregular Systems" (Ph. Blanchard et al., eds.). World Scientific, 1993.
17
[31]� Symanzik, K.: Euclidean quantum field theory. In "Local Quantum Theory" (R.� Jost, ed.). Academic, New York (1969).
[32]� Varadhan, S. R. S.: Appendix to "Euclidean quantum field theory" by K. Symanzik. In "Local Quantum Theory" (R. Jost. ed.). Academic. New York (1969).
[33]� Watanabe, H.: The local time of self-intersections of Brownian motions as generalized Brownian junctionals. Lett. Math. Phys. 23, 1-9, 1991.
[34]� Werner, W.: Sur les singularites des temps locaux d'intersection du mouvement brownien plan. Ann. Inst. H. Poincare 29 (1993) 391-418.
[35]� Westwater, J.: On Edward's model for long polymer chains. Cornm. Math. Phys. 72 (1980), 131-174.
[36]� Wolpert, R.: Wiener path intersection and local time. J. FUnct. Anal. 30 (1978), 329-340.
[37]� Yor, M.: Complements aux formules de Tanaka-Rosen. Sem. de Prob. XIX, 1983/84. LNM 1123 (1985), 332-348. Springer, Berlin.
[38]� Yor, M.: Renormalisation et ~onvergenc~ en, lo~ p~ur les temps l.o~ux d'intersection du mouvement brownten dans R . Semmarre de Probabllite, 1983/84. LNM 1123 (1985), 350-365. Springer, Berlin.
18�