I L NUOVO CIMENTO VOL. VIII, N.

2 16 Aprile 1958

On the Solution of Certain Singular Integral Equations

of Quantum Field Theory.

1~. OM~s (*)

CERN - Geneva

(ricevuto il 19 Gennaio 1958)

Summary. - - In the study of complex phenomena involving n-mesons

(e.g. simple and double photoproduction, radiative scattering, double
production, etc.) one encounters singular integral equations linking the
matrix elements and phase-shifts. We give here a general method of
resolution of these equations which furnishes the following results: a) the
simplest type is integrable by quadratures; b) the general type is reducible
to a Fredholm equation. We give also the solution of a system of coupled
integral equations and of the particular one which occurs in the double
production problem. This last case may be integrated by quadratures.

1. - Introduction.

W e w a n t to s t u d y singular integral equations of the following t y p e

co

~1.1)
! fh*(x')~(~')d x ' d - fg (x'x)q)(x')dx/,
~s(x) = l(x) d- 7 ~ J x ' - - x - - i s
s

w h e r e h(x) has the f o r m ei~sin d (real d) a n d K ( x x ' ) is a regular kernel.

This t y p e of equation f r e q u e n t l y occurs in the static t h e o r y of Chew-Low-
Wick (1,2) for instance w h e n one studies processes initiated or ending b y a
=-nucleon system. Such equations are thus f u n d a m e n t a l in the p r o b l e m s of
,simple (1) and double p h o t o p r o d u c t i o n , r a d i a t i v e scattering (3), p r o d u c t i o n of

(*) On leave of absence from CEA, Saelay, France.

(1) G. F. Cm~w and F. E. Low: Phys. Rev., i0I, 1579 (1956).
(2) C. G. WICK: Rev. Mod. t)hys., 27, 339 (1955).
(a) B. Bosco: 3[uovo Cime~to, 5, 1361 (1957).
 ON T H E S O L U T I O N O F C E R T A I N S I N G U L A R I N T E G R A L E Q U A T I O N S :ETC. 317

mesons b y meson (4.6). However, we suppose t h a t the interest of such ~

problem is not completely restricted to the static approximation.
E q u a t i o n (1.1) is not of F r e d h o l m t y p e although analogous in form and~
for instance, classical results such as F r e d h o l m alternative are not applicable.
Moreover, due to the kernel singularity, these equations do not l e n d t h e m -
selves simply to approximation methods and numerical integration. This p o i n t
does not seem to have been clearly stressed in literature and errors in t h e
choice of approximation methods have been made concurrently (ef. ref. (e)).
As the mathematical problem seemingly will keep its importance, we w a n t
to give to it in this article a form upon which the classical methods of ap-
proximation and the general theorems will be applicable. B y the way, we
shall be able to show in particular cases the existence of a solution and its
non-uniqueness, and shall determine what conditions of a physical n a t u r e allow
complete uniqueness. I n fact the m e t h o d can be generalized to a quite wide
v a r i e t y of problems which we shall indicate.
I n Sect. 2 we give the general solution of the simplest equation of t y p e (1.1).
We s t u d y the uniqueness problem when certain conditions of physical n a t u r e are
imposed. Sect. 3 deals with the reduction of a quite general case to Fredholm's
one; the problem of the equivalence between these two forms is not treated.
Sect. 4 generalizes this reduction to systems of coupled integral equations.
Finally, we t r e a t in Sect. 5 a particular system of coupled equations which
is encountered in the problem of 7:-meson production b y mesons (~.e). This
last case is seen to reduce to simple quadratures, just as the problem of Sect. 2.

2. - Fundamental equation.

The simplest equation of t y p e (1.1) is

cJr

(2.1) 1 f~*(x')v(x')
~(x) = J(x)~ 4- ~ j x , _ x _ i e d x ' ,
1

here ](x) is a given function which we suppose bounded between 1 and c~.
h(x) is a given function of the form exp [i~] sin (~, where ~(x) is bounded.
Whe choose arbitrarily the determination of ~ which tends to 0 at infinity
(supposing this choice possible) and ~ ( 1 ) = kz. ~ tends to zero at least as
quickly as x -1 when x tends to infinity. ~(x) is the u n k n o w n function. I n
order to solve (2.1) we shall use a m e t h o d which is a generalization of t h e

(~) J. FUKUDA and J. S. KOVACS: Phys. Rev., 104, 1784 (1956).

(5) L. S. RODBERG: Phys. Rev., 106, 1090 (1957).
(8) R. OMNgS: Nuovo Cimento, 6, 780 (1957).
318 ~. OMN~S

m e t h o d introduced b y N. I. MVSX_TIELIS~Vn~I(7) for t h e solution of I t f l b e r t

p r o b l e m in elasticity, the essential difference being t h a t here the kernel c o n -
rains h(x) and is not a simple Cauchy kernel. W e shall h a v e to introduce
functions of a v a r i a b l e z defined in a complex plane cut along the i n t e r v a l
(1, oc) a n d shall call G(x+) (resp. G(x--)) the limit of a function G(z) when
z tends to x u p o n (resp. under) the cut. I n the following, the cut will be
called L.
L e t us define the function

(2.2) ~(z) = ~ f
L
x'-- z dx,,
which implies

(2.3a)
1 lh*(x'!v(x ')
:~j x ' - - x -- is dx'---- 2iF(x ~- ) ,
L

~ ~--I
(2.3b) ~(x) h,~, [F(x + ) " F ( x - - ) ] .

E q u a t i o n (2.1) takes now the f o r m

(2.~) exp [-- 2i(~]F(x ~-) - - F ( x - - ) = ](x)h*(x),

where we h a v e used 1 - 2 i h * = exp [--2i~]. L e t us now p u t

(2.5) F(z) = O ( z ) 9 ( z ) ,

where the function ~Q(z) is defined b y the condition

(2.6) exp [ - - 2 i ( ~ ] ~ Q ( x - ~ ) - tg(x--) = O.

This last equation admits a solution (which is found b y taking the l o g a r i t h m

of (2.6))
~9(z) = exp [u(z)]
with

= de,
L

if we define
co

(2.7) l~f ~($) d~

1

(7) MUSKHELISI~VlLI:Trud. Tbil. Mat. Inst., 10, 1 (1941), cited in S. G. MIKHLIN:

Integral Equations (London, 1957), p. 126 ft.
 ON THE SOLUTION OF CERTAIN SINGULAR INTEGRAL EQUATIONS ETC. 319

we have

,(2.8) 9(x + ) = exp [q + iS]; X2(x--) = exp [q'--iS].

Equation (2.4) m a y now be transformed ~in a relation for ~5(z) which is

,of the ttflbert t y p e

,(2.9) qS(x ~- ) -- q~(x--) = ](x)h*(x)~-~(x - ) = ](x) sin S(x) exp [-- ~(x)]',

which has a solution

1 fJ(~) sin 3(~) exp [-- ~(~)] de:

,(2.1o/ qS(z) = ~ ~ -- z
L

.one m a y now derive ~(x) by (2.1) or (2.3b) which give evidently the same result

<2.11) q~(x) = [/(x) cos S(x)~ + x-1exp [q(x)].

co

exp [i S(x)].
l

I t is an easy task to verify t h a t every step of this method is correct, provided

t h e written integrals converge and t h a t (2.11) is truly a solution of (2.1).
I t is important to point out t h a t (2.11) is not the only solution of (2.1)
~or one m a y add to it any solution of the homogeneous equation

1 fh*(x')q~(x')
<2.12) %(x) = ~j x'-- x ~ e dx',
L

;such a solution of (2.12) m a y y e t be defined by (2.3) and a relation analogous

to (2.5)

<2.13) Fo(Z) = r

where ~o(Z) must now verify

<2.1~) r + ) -- r = o.

This last relation shows t h a t ~50(z) is an analytic function in the whole

complex plane except eventually at the points I and ~ where it m a y have
singularities. If we exclude essential singularities, the general solution of (2.1)
320 n. OMN~S

a p p e a r s as
qJ(x) = of(x) + L(x) exp [e(x) + i h ( x ) ] ,
(2.15) P(x)
L(x) (x-- 1},,'

where P(x) a n d n are a r b i t r a r y p o l y n o m i a l and integer. Generally a solution

will be completely d e t e r m i n e d b y its a s y m p t o t i c b e h a v i o u r in t h e neigh-
b o u r h o o d of 1 and co for this b e h a v i o u r b y (2.11) and (2.15) d e t e r m i n e s n
a n d P(x). P a r t i c u l a r l y , for

a) k = 0 or 5(1) = O,

b) k------1 or 5(1) = - - z ; iS(s) + z ] ~ J /~>1,

(2.11) is the only solution of (2.1) regular in the neighbourhood of 1 a n d w h i c h

tends to 0 at infinity (see appendix). These cases are precisely the m o r e in-
teresting in practical use since, in the c o n t e x t of Chew a n d L o w model, case a)
m a y be a d o p t e d for the small phase-shifts a n d case b) for 58 w i t h fl = 3. L e t
us recall here t h a t we h a v e t a k e n the d e t e r m i n a t i o n of 25 which is 0 a t infinity.
The choice k = - - 1 is equivalent to suppose only one resonance for t h e (~3~)
pion-nucleon state. L e t us stress t h a t h y p o t h e s e s a) a n d b) simplify h a p p i l y
the calculations b u t are absolutely not essential to the success of this approach.
If, effectively, a) and b) are not verified b y 5, it is an easy t a s k to d e t e r m i n e
the singularities of the integrals a p p e a r i n g in (2.7) and (2.11) using the first
t e r m s of the T a y l o r expansion of 5 n e a r 1, a n d to choose in a unique fashio~
an integer n a n d a p o l y n o m i a l P(x) in (2.15) in order to h a v e a regular solution
in the neighbourhood of 1 and c~. W e shall consider this solution as the in-
teresting one although we do n o t w a n t to enter here in the difficult p r o b l e m
of the choice of physical criteria for a solution (see ref. (s-10)).
L e t us n o w r e m a r k t h a t , for ](x) real, the p h a s e of ~(x) is (~ which is inti-
m a t e l y connected with a t h e o r e m b y FUBINI~ ~I-A1VfBUand WATAGm~ (11).
Finally, in the m o r e general case where h has n o t the f o r m e~sin 5 o u r
m e t h o d is y e t applicable in principle, the essential condition being n o w t h a t
[1 - - 2ih*[ m u s t n o t be zero on L. One m a y y e t write 1 - - 2ih* = exp [-- 2iy],
where y is n o w complex. The m e t h o d used b y MUSKttELISHVILI (~) in the re-
solution of H i l b e r t ' s p r o b l e m is a p p a r e n t l y ~ p a r t i c u l a r case (*).

(s) •. CASTILLEJO, R. H. DALITZ and F. J. DYsoN: Phys. Rev., 101,453 (1956).

(9) F. J. DYsoN: Phys. Rev., 106, 157 (1957).
(lo) R. HAAG: NUOVO Cimento, 5, 203 (1957).
(11) S. FV]3INI, Y. NAMer and V. WATAG•IN: to be published.
(*) s added in proo]: This general ease was known to MusxnELISHVlLI and
treated in his book "Singular Integral Equations ", Groningen, 1953; this was kindly
pointed out to us by J. LAscoux.
 ON THE SOLUTION OF CERTAIN SINGULAR INTEGRAL EQUATIONS ETC. 321

3. - E q u a t i o n reducible to Fredholm's type.

L e t us consider t h e equation

co
(3.1)
f[ h,(x,)
q~(x) = ](x) + ~ J [ x ' - - x - - ii
]
+ X ( x ' x ) ~(x') dx',
1
where the kernel K(xx') is regular. Including fKq~ in the inhomogeneous part~
one finds a solution analogous to (2.11) where ](x) is n o w replaced b y

](x) -+-l f K(x'x)cf(x')dx',

if we suppose q(x) b o u n d e d and continuous a n d t h a t K(xx') verifies K(xx') ~ 0

when x - + c% and is s u b m i t t e d to a Lipshitz condition, one m a y i n v e r t t h e
order of integrations in

P ~j~ .K(x'x)~(x') dx',

(3.2)
2(~) = sin 5($) exp [ - - e(~)] ,

which leads to

(3.3) v(x) = + JN(x'x)v(x')dx',

1
where

(3. ~) [~(x)-----[j(x) cOs(~(x)-j[-lex r / x \ T P ~](~)~(~) 1

1

(3.5)

H e r e we shall n o t t r y to determine in detail w h a t are necessary conditions

for K and h in order t h a t e q u a t i o n (3.3) be of F r e d h o l m t y p e : we think it t o
be a m a t t e r of interest only in each p a r t i c u l a r case. The solution we h a v e
o b t a i n e d m u s t evidently be regular at i a n d c~ a n d one m a y r e p e a t here t h e
a r g u m e n t s given in the preceding section.
~22 R. OMN]~S

4 . - The case of a s y s t e m of coupled integral equations.

L e t us consider the system

co

'(~,1) ~)i(X) = ]i(X) Jr-

~Jf[h~i(xl)~i(x
[x~ t) -- ]
-~- ~ ~ij(XrX)Cf;(X f) d x ' ,
1

b y the m e t h o d of the Sect. 3, this m a y be p u t eventually in F r e d h o l m form b y

,(4.2) ,(x' x)q~j(x') dx',

T
"where

(4.3) J/#~(x) =
I,
, cos 5i + ~ exp [p~]
q'xX: 1
~( )2~_ ) dx' exp [i0,(x)],

Nif(XtX) = [K,(x'x) cos 6~(x') + U~j(x'x)] exp [ i ~ d x ) ] ,

(4.4) V.(x'x)= ex P ~Le~(x)~P

J Jf ~ . ( x ' ~ ) 2 d ~ )

5 . - E q u a t i o n s for the production of m e s o n s by m e s o n s .

W h e n one studies the reaction n u c l e o n + ~ -~ nucleon-~2~ b y the methods

of Chew-Low-Wick, one obtains equations of the following t y p e for reduced
:matrix elements (*)
co
1 ['h$i(x')~i(X'XlX2)
t5.1) q~i(xx~x2) = ]~(xxlx2) + ~ j x~-x~ ~ dx' ~-
1
8
+ ~A,G~(xlx2), (i = 1, ..., S)i

"where A~ are real numbers and

(5.2) r = *(x'xlx2)hj(x') x ' - - xl - - is + x ' - - x~ - - is '

(*) These equations reproduce equations (5.1) of reference (s) where we have put,
in order to save writing ~i = Ts~ and we have taken into account relation Ts~ = (--)2STsL
~which follows by time-reversal invariance.
 ON T H E $OLUTION OF CERTAIN SINGULAR I N T E G R A L E Q U A T I O N $ ETC. 323

equation (5.1) is in fact a partieulur case of (2.1) ~nd its solution is given by

(5.3) ~(xx, x~) = ~>(xx~x~) + q~(~(XX~X~),

where

,(5.4) ~'2(xxlx~) = []~(xxlx~) cos ai(x) + x-1exp [e,(z)].

co
t )x;2)~i(~)d~/91 exp [i ~(x)]
" j ~]i(~xl
l

.(5.5) q~i2)(xxlx2) = ~ A,Cj(xlx~)ui(x) ,

J

~f we define the operations Di[~f(XXlX2)] which transforms a function ~ of

(x, x~, x2) in ~ function of x~ and x2 only by

(5.7) Di[y~] : *(x%x2)h~(x') - ~ x ' - - x 2 - - i e dx',

(5.s) D~[~] = C~(x~x~)

und apply it to (5.4), we obtain

(5.9) Ci(xix2) : ~Di[~i1)] + ~ A,C*D~[u~].

In (5.9) D~[~'~~)] and D~[u~] may be explieitely calculated and one may easily
solve (5.9) for the Cr and bring them in (5.3-6). It is seen that the solution
so obtained involves only quadratures, which is indeed an unexpected simple
result.

6. - Conclusions.

We have given a method for the resolution of integral equations which

present themselves in quantum field theory and, in particular, in the model
of Chew-Low and Wick. Following the difficulties involved in the considered
324 ~. OMN~S

problem, one is led to an explicit solution b y quadratm'es or to non-singular

integral equations. The m e t h o d m a y be generalized to a quite wide lot of
other eases. An explicit and suggesting example is given b y the equstions for
production of mesons.

* $ *

The a u t h o r is h a p p y to t h a n k Professor C. J. ]3AKKEI%for his hospitality

at CEI%N and Professor B. ~ERI%ETTI :[or his constant interest.

APPENDIX

I n this appendix, we want to s t u d y the convergence of integrals appearing

in equations (2.8) and (2.11). I n this respect, let us consider

co

(a.1) = P (:) d : .
j~--x
o

Here we have replaced the limit 1 b y 0 in order to simplify developments

and /~(~) is a b o u n d e d and derivable function such t h a t

(A.2) F(~) = $~ + 0 , ~ > Z, ~ > 0,

(A.3) ~ ( ~ ) = k + B ~ p + 0(~ z) , ~ ~ 0, ~ > 0,

L e t us s t u d y I(x) for x near to 0 and infinitely great.

a) x great.
L e t us p u t
z

(A.4) I = I1 + I5 I1 = f ~ ( ~ ) d~
~ - - X
o

co

for x >> ZI1 is of order const/x + O(1/x). If we define ~ ---- n + s (0 < s 1)

 ON T H E S O L U T I O N OF C E R T A I N S I N G U L A R I N T E G t C A L E Q U A T I O N S E T C . 325

it comes
~ (-)"-~+1
(A.5) 12 = A _ xn-~+~Z ~+~'-~(p + s 1) +
c~

p (O(1/~) d~ .
+ (--)"+~x" ~"( x) d ~ 4 - J$-x
z, 7,

:By t h e b o u n d e d n e s s of t h e d e r i v a t i v e t h e l a s t i n t e g r a l is 0 @ -1) a n d t h e p r e -
c e d i n g o n e is b o u n d e d b y A x -n l o g ( x - Z ) Z -~' l o g Z. T h e r e s u l t s of T a b l e I
follow.
TABLE I.

~, X Cases Order of I(x) Order of exp [/(x)]

oo ~>1 x -1 1
= 1 x -1 log x 1
~<1 ? ?

0 k :/: 0 - - k log x x -~
k=O 1 1

b ) x near to O.
T h e m e t h o d is a n a l o g o u s , o n e uses (A.3) a n d s e p a r a t e s t h e p a r t s of I d u e
t o k, t o f a n d 0 ( ~ ) , w h i c h gives, t e r m - b y - t e r m

I ( x ) = - - k l o g x + c o n s t + O(x) ,

f r o m w h i c h r e s u l t s of T a b l e I follow.
I n T a b l e I , w e g i v e t h e a s y m p t o t i c v a l u e s of I ( x ) a n d e x p [ I ( x ) ] . Table II

TABLE I I .

Cases Order of Order of eQ Order of v Order of J

~z>l 1
~=1 1
~<1 ?

k#O 1 X-k ~+~ ?

fi< --k
fi=--~ 1 X k log x
1 X k 1
k=0 1 1
326 R. o M ~ s

is a direct application of the results of T a b l e I to ~(~) and v(~) ----](~) sin ~(~)-
9exp [ - - ~(~)] where ](~) is bounded, a n d finally of J + ----PSv(~)d~/(~ - - x).
One sees, as i n d i c a t e d in the t e x t , t h a t if fi ~ - k~ t h e ' i n t e g r a l of (2.11)
converges a n d if k ~ 0 exp [Q(x)] is e v e r y w h e r e bounded.

RIASSUNTO (*)

Nello studio dei s complessi interessanti i mesoni :: (ad es. fotoproduzione

semplice e doppia, scattering radiativo, produzione doppia, eee) si ineontrano equa-
zioni integrali che collegano gli elementi di matrice col spostamenti di fase. Diam(~
qui un metodo generale per la soluzione di queste equazioni ehe conduce ai seguenti
risultati: a) il tipo pifi sempliee ~ integrabfle per quadrature; b) il tipo generale
riducibfle a una equazione di Fredholm. Diamo anche la soluzione di un sistema di
equazioni integrali aceoppiate e di quella particolare equazione che interviene nel pro-
blema della produzione doppia. Quest'ultimo easo pus essere integrato per quadrature.

(*) T r a d u z i o n e a cura della R e d a z i o n e .