The sums of squares function (Engl. Sum of squares function ) is a number theoretic function that specifies how many ways a given natural number as a sum of can be represented squares of integers, all permutations, and sign combinations are counted. It is referred to as .
n
{\ displaystyle n}
k
{\ displaystyle k}
r
k
(
n
)
{\ displaystyle r_ {k} (n)}
definition
The first values of r k
n
n
r 1 (n)
r 2 (n)
r 3 (n)
r 4 (n)
r 5 (n)
r 6 (n)
r 7 (n)
r 8 (n)
0
0
1
1
1
1
1
1
1
1
1
1
2
4th
6th
8th
10
12
14th
16
2
2
0
4th
12
24
40
60
84
112
3
3
0
0
8th
32
80
160
280
448
4th
2 2
2
4th
6th
24
90
252
574
1136
5
5
0
8th
24
48
112
312
840
2016
6th
2‧3
0
0
24
96
240
544
1288
3136
7th
7th
0
0
0
64
320
960
2368
5504
8th
2 3
0
4th
12
24
200
1020
3444
9328
9
3 2
2
4th
30th
104
250
876
3542
12112
10
2‧5
0
8th
24
144
560
1560
4424
14112
11
11
0
0
24
96
560
2400
7560
21312
12
2 2 ‧3
0
0
8th
96
400
2080
9240
31808
13
13
0
8th
24
112
560
2040
8456
35168
14th
2‧7
0
0
48
192
800
3264
11088
38528
15th
3‧5
0
0
0
192
960
4160
16576
56448
16
2 4
2
4th
6th
24
730
4092
18494
74864
17th
17th
0
8th
48
144
480
3480
17808
78624
18th
2‧3 2
0
4th
36
312
1240
4380
19740
84784
19th
19th
0
0
24
160
1520
7200
27720
109760
20th
2 2 ‧5
0
8th
24
144
752
6552
34440
143136
The function is defined as
r
k
(
n
)
: =
∑
a
1
2
+
a
2
2
+
⋯
+
a
k
2
=
n
(
a
1
,
a
2
,
...
,
a
k
)
∈
Z
k
1
=
|
{
(
a
1
,
a
2
,
...
,
a
k
)
∈
Z
k
∣
a
1
2
+
a
2
2
+
⋯
+
a
k
2
=
n
}
|
{\ displaystyle r_ {k} (n): = \ sum _ {\ begin {array} {c} a_ {1} ^ {2} + a_ {2} ^ {2} + \ cdots + a_ {k} ^ {2} = n \\ (a_ {1}, a_ {2}, \ dots, a_ {k}) \ in \ mathbb {Z} ^ {k} \ end {array}} 1 \ quad = \ quad { \ big |} \ left \ {(a_ {1}, a_ {2}, \ dots, a_ {k}) \ in \ mathbb {Z} ^ {k} \ mid a_ {1} ^ {2} + a_ {2} ^ {2} + \ cdots + a_ {k} ^ {2} = n \ right \} {\ big |}}
,
d. H. as the number of possible representations of as the sum of squares of whole numbers with . Be it .
n
{\ displaystyle n}
k
{\ displaystyle k}
k
≥
1
{\ displaystyle k \ geq 1}
r
k
(
0
)
: =
1
{\ displaystyle r_ {k} (0): = 1}
For example , because there is always 2 sign combinations, and also because of 4 sign combinations. On the other hand, it is because there is no representation of the number 3 as the sum of 2 squares.
r
2
(
1
)
=
4th
{\ displaystyle r_ {2} (1) = 4}
1
=
0
2
+
(
±
1
)
2
=
(
±
1
)
2
+
0
2
{\ displaystyle 1 = 0 ^ {2} + (\ pm 1) ^ {2} = (\ pm 1) ^ {2} + 0 ^ {2}}
r
2
(
2
)
=
4th
{\ displaystyle r_ {2} (2) = 4}
2
=
(
±
1
)
2
+
(
±
1
)
2
{\ displaystyle 2 = (\ pm 1) ^ {2} + (\ pm 1) ^ {2}}
r
2
(
3
)
=
0
{\ displaystyle r_ {2} (3) = 0}
The relationship immediately follows from the definition
r
k
+
m
(
n
)
=
∑
t
=
0
n
r
k
(
t
)
r
m
(
n
-
t
)
,
{\ displaystyle r_ {k + m} (n) = \ sum _ {t = 0} ^ {n} r_ {k} (t) \ r_ {m} (nt),}
from which a recursion formula for efficient calculation can be derived:
r
k
+
1
(
n
)
=
r
k
(
n
)
+
2
∑
t
=
1
n
r
k
(
n
-
t
2
)
.
{\ displaystyle r_ {k + 1} (n) = r_ {k} (n) +2 \ sum _ {t = 1} ^ {\ sqrt {n}} r_ {k} (nt ^ {2}). }
Average order of magnitude
Be it
R.
k
(
x
)
: =
∑
n
=
0
x
r
k
(
n
)
=
∑
a
1
2
+
a
2
2
+
⋯
+
a
k
2
≤
x
1
{\ displaystyle R_ {k} (x): = \ sum _ {n = 0} ^ {x} r_ {k} (n) = \ sum _ {a_ {1} ^ {2} + a_ {2} ^ {2} + \ dotsb + a_ {k} ^ {2} \ leq x} 1}
.
This is clearly the number of (integer) grid points in a -dimensional sphere with the radius and therefore approximately equal to the sphere volume. More precisely, it can be derived recursively
k
{\ displaystyle k}
x
{\ displaystyle {\ sqrt {x}}}
R.
k
(
x
)
=
V
k
x
k
2
+
O
(
x
k
-
1
2
)
{\ displaystyle R_ {k} (x) = V_ {k} x ^ {\ frac {k} {2}} + O (x ^ {\ frac {k-1} {2}})}
,
where the constants are the volumes of the -dimensional unit spheres: and is the Landau symbol .
V
k
{\ displaystyle V_ {k}}
k
{\ displaystyle k}
V
2
=
π
,
V
3
=
4th
3
π
,
V
4th
=
1
2
π
2
,
...
{\ displaystyle V_ {2} = \ pi, \; V_ {3} = {\ tfrac {4} {3}} \ pi, \; V_ {4} = {\ tfrac {1} {2}} \ pi ^ {2}, \; \ dots}
O
(
.
)
{\ displaystyle O (.)}
The average order of magnitude of is thus , e.g. B. that of .
r
k
(
n
)
{\ displaystyle r_ {k} (n)}
k
2
V
k
x
k
2
-
1
{\ displaystyle {\ tfrac {k} {2}} V_ {k} x ^ {{\ tfrac {k} {2}} - 1}}
π
{\ displaystyle \ pi}
r
2
(
x
)
{\ displaystyle r_ {2} (x)}
Generating function
The generating function is obtained as the power of the Jacobian theta function for the special case . This is true
ϑ
(
z
,
q
)
{\ displaystyle \ vartheta (z, q)}
z
=
0
{\ displaystyle z = 0}
ϑ
3
(
q
)
: =
ϑ
(
0
,
q
)
=
∑
n
=
-
∞
∞
q
n
2
=
1
+
2
q
+
2
q
4th
+
2
q
9
+
2
q
16
+
⋯
{\ displaystyle \ vartheta _ {3} (q): = \ vartheta (0, q) = \ sum _ {n = - \ infty} ^ {\ infty} q ^ {n ^ {2}} = 1 + 2q + 2q ^ {4} + 2q ^ {9} + 2q ^ {16} + \ dotsb}
One gets from it
(
ϑ
3
(
q
)
)
k
=
∑
n
1
,
n
2
,
...
,
n
k
q
n
1
2
+
n
2
2
+
⋯
+
n
k
2
=
∑
n
=
0
∞
q
n
∑
n
1
2
+
n
2
2
+
⋯
+
n
k
2
=
n
1
=
∑
n
=
0
∞
q
n
r
k
(
n
)
{\ displaystyle (\ vartheta _ {3} (q)) ^ {k} = \ sum _ {n_ {1}, n_ {2}, \ dotsc, n_ {k}} q ^ {n_ {1} ^ { 2} + n_ {2} ^ {2} + \ dotsb + n_ {k} ^ {2}} = \ sum _ {n = 0} ^ {\ infty} q ^ {n} \ sum _ {n_ {1 } ^ {2} + n_ {2} ^ {2} + \ dotsb + n_ {k} ^ {2} = n} 1 = \ sum _ {n = 0} ^ {\ infty} q ^ {n} \ r_ {k} (n)}
Special cases
Values and average order of magnitude of r
2 (n)
Values and average order of magnitude of r
4 (n)
Values and average order of magnitude of r
8 (n)
Simple formulas result for straight lines , e.g. B. ( ):
k
{\ displaystyle k}
n
>
0
{\ displaystyle n> 0}
The following applies to:
k
=
2
{\ displaystyle k = 2}
r
2
(
n
)
=
4th
∑
d
∣
n
d
≡
1
(
mod
2
)
(
-
1
)
(
d
-
1
)
/
2
{\ displaystyle r_ {2} (n) = 4 \ sum _ {d \ mid n \ atop d \ equiv 1 {\ pmod {2}}} (- 1) ^ {(d-1) / 2}}
With the help of prime factorization , where the prime factors are the form and the prime factors are the form , another formula results
n
=
2
G
p
1
f
1
p
2
f
2
⋯
q
1
H
1
q
2
H
2
⋯
{\ displaystyle n = 2 ^ {g} p_ {1} ^ {f_ {1}} p_ {2} ^ {f_ {2}} \ cdots q_ {1} ^ {h_ {1}} q_ {2} ^ {h_ {2}} \ cdots}
p
i
{\ displaystyle p_ {i}}
p
i
≡
1
(
mod
4th
)
{\ displaystyle p_ {i} \ equiv 1 {\ pmod {4}}}
q
i
{\ displaystyle q_ {i}}
q
i
≡
3
(
mod
4th
)
{\ displaystyle q_ {i} \ equiv 3 {\ pmod {4}}}
r
2
(
n
)
=
4th
(
f
1
+
1
)
(
f
2
+
1
)
⋯
{\ displaystyle r_ {2} (n) = 4 (f_ {1} +1) (f_ {2} +1) \ cdots}
,
when all exponents are even. If at least one is odd, then is . By definition, the number of all Gaussian numbers is also with the norm .
H
1
,
H
2
,
...
{\ displaystyle h_ {1}, h_ {2}, \ dotsc}
H
i
{\ displaystyle h_ {i}}
r
2
(
n
)
=
0
{\ displaystyle r_ {2} (n) = 0}
r
2
(
n
)
{\ displaystyle r_ {2} (n)}
n
{\ displaystyle n}
The formula for comes from Carl Gustav Jacob Jacobi and is given as the eightfold sum of all factors that are not divisible by 4 ( Jacobi's theorem ):
k
=
4th
{\ displaystyle k = 4}
n
,
{\ displaystyle n,}
r
4th
(
n
)
=
8th
∑
d
∣
n
;
4th
∤
d
d
{\ displaystyle r_ {4} (n) = 8 \ sum _ {d \ mid n; 4 \ nmid d} d}
r
4th
(
n
)
{\ displaystyle r_ {4} (n)}
is also the number of all Lipschitz quaternions with the norm .
n
{\ displaystyle n}
Jacobi also found an explicit formula for :
k
=
8th
{\ displaystyle k = 8}
r
8th
(
n
)
=
16
∑
d
∣
n
(
-
1
)
n
+
d
d
3
{\ displaystyle r_ {8} (n) = 16 \ sum _ {d \ mid n} (- 1) ^ {n + d} d ^ {3}}
Relationship to the Sierpiński constant
The Limes
K
: =
lim
n
→
∞
(
∑
k
=
1
n
r
2
(
k
)
k
-
π
log
n
)
{\ displaystyle K: = \ lim _ {n \ to \ infty} \ left (\ sum _ {k = 1} ^ {n} {\ frac {r_ {2} (k)} {k}} - \ pi \ log n \ right)}
exists and is called (after Wacław Sierpiński ) the Sierpiński constant . This can be expressed by the number of circles , the Euler-Mascheroni constant and the gamma function :
K
=
π
(
2
γ
+
4th
log
Γ
(
3
4th
)
-
log
π
)
{\ displaystyle K = \ pi (2 \ gamma +4 \ log \ Gamma ({\ tfrac {3} {4}}) - \ log \ pi)}
See also
Web links
Individual evidence
↑ E. Krätzel: Number theory . VEB Deutscher Verlag der Wissenschaften, Berlin 1981, p. 165 .
↑ E. Krätzel: Number theory . VEB Deutscher Verlag der Wissenschaften, Berlin 1981, p. 197 .
<img src="https://de.wikipedia.org/wiki/Special:CentralAutoLogin/start?type=1x1" alt="" title="" width="1" height="1" style="border: none; position: absolute;">