OEIS/Eta products: Difference between revisions
imported>Gfis Normalization, Exponentiation, Multiplication |
imported>Gfis corr. |
||
Line 3: | Line 3: | ||
Following Michael Somos, we describe an eta product by a matrix resp. a list of pairs ''(q<sub>i</sub>, e<sub>i</sub>)'' of the form | Following Michael Somos, we describe an eta product by a matrix resp. a list of pairs ''(q<sub>i</sub>, e<sub>i</sub>)'' of the form | ||
[q1,e1;q2,e2;q3,e3;...] | [q1,e1;q2,e2;q3,e3;...] | ||
The ''q<sub>i</sub>'' are the powers of ''q'' and the ''e<sub>i</sub>'' are the powers of the eta functions. | The pairs are separated by ";". | ||
* The ''q<sub>i</sub>'' are the powers of the argument ''q'' inside the eta function , and | |||
* the ''e<sub>i</sub>'' are the powers of the eta functions. Only these can be negative. | |||
The following section contains a number of examples. | |||
===Common generating functions=== | ===Common generating functions=== | ||
====Partition numbers==== | ====Partition numbers==== | ||
Line 60: | Line 63: | ||
For long lists of generating functions it is convenient to require all EPSs to be normalized, such that they can be tested for equality and can be sorted. | For long lists of generating functions it is convenient to require all EPSs to be normalized, such that they can be tested for equality and can be sorted. | ||
We use the following rules: | We use the following rules: | ||
# The sublist of pairs with positive ''e<sub>i</sub>'' comes before the sublist with negative ''e<sub>i</sub>''. | |||
# Inside a sublist the first components ''q<sub>i</sub>'' are in descending order. | |||
# Successive pairs with the same ''q<sub>i</sub>'' are combined into a single pair with the sum of the ''e<sub>i</sub>'' as second component. | |||
===Exponentiation of an EPS=== | ===Exponentiation of an EPS=== | ||
An EPS is taken to a power ''m'' by simply multiplying all ''q<sub>i</sub>'' (the first components in the pairs) by ''m''. The exponent may be ''1/2'' for the square root, or any other fraction ''m<sub>n</sub>/m< | An EPS is taken to a power ''m'' by simply multiplying all ''q<sub>i</sub>'' (the first components in the pairs) by ''m''. The exponent may be ''1/2'' for the square root, or any other fraction ''m<sub>n</sub>/m<sub>d</sub>'' as long as the ''q<sub>i</sub>'' are all divisible by ''m<sub>d</sub>''. | ||
===Multiplication of two EPSs=== | ===Multiplication of two EPSs=== | ||
Two EPSs could be multiplied by simply concatenating their list of pairs. It is convenient, howver, to combine all occurrences of specific ''q<sub>i</sub>'' | Two EPSs could be multiplied by simply concatenating their list of pairs. It is convenient, howver, to combine all occurrences of specific ''q<sub>i</sub>'' | ||
by adding the corresponding eta exponents ''e''. | by adding the corresponding eta exponents ''e''. |
Revision as of 22:04, 23 January 2023
Eta product signatures (EPS)
The coefficient sequences for generating functions that are products of the eta function can be computed conveniently by an Euler transform of some periodic integer sequence. Following Michael Somos, we describe an eta product by a matrix resp. a list of pairs (qi, ei) of the form
[q1,e1;q2,e2;q3,e3;...]
The pairs are separated by ";".
- The qi are the powers of the argument q inside the eta function , and
- the ei are the powers of the eta functions. Only these can be negative.
The following section contains a number of examples.
Common generating functions
Partition numbers
A000041 1, 1, 2, 3, 5, 7, 11, 15, 22, 30, 42, 56, 77, 101, 135, 176, 231 Euler transform of period 1: [1]
Dedkind eta η (without the q^(1/24) factor)
A010815 1, -1, -1, 0, 0, 1, 0, 1, 0, 0, 0, 0, -1, 0, 0, -1, 0, 0, 0, 0, 0, 0, 1 eta(q) Euler transform of period 1: [-1] eps P="[1,1]"
Jacobi theta_2 θ_2
A089800 2, 0, 2, 0, 0, 0, 2, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 2 ... = theta_2(q)/q^(1/4)
Jacobi theta_3 θ_3, Ramanujan phi φ
A000122 1, 2, 0, 0, 2, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 2 a(0) = 1, for n >= 1: a(n) = 2 if n is a square, otherwise 0. eta(q^2)^5 / (eta(q)*eta(q^4))^2 Euler transform of period 4: [2,-3,2,-1] eps P="[2,5;1,-2;4,-2]"
Jacobi theta_4 θ_4
A002448 1, -2, 0, 0, 2, 0, 0, 0, 0, -2, 0, 0, 0, 0, 0, 0, 2 a(0) = 1, for n >= 1: a(n) = 2 * (-1)^sqrt(n) if n is a square, otherwise 0. eta(q)^2 / eta(q^2) Euler transform of period 2: [2,-1] eps P="[1,2;2,-1]"
Ramanujan psi ψ
A010054 1, 1, 0, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0 a(n) = 1 if n is a triangular number, otherwise 0. q^(-1/8) * eta(q^2)^2 / eta(q) Euler transform of period 2: [1,-1] eps P="[2,2;1,-1]"
Ramanujan chi χ
A000700 1, 1, 0, 1, 1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 4, 5 q^(1/24) * eta(q^2)^2 /(eta(q) * eta(q^4)) Euler transform of period 4: [1,-1,1,0] eps P="[2,2;1,-1;4,-1]"
Ramanujan f
A121373 1, 1, -1, 0, 0, -1, 0, -1, 0, 0, 0, 0, -1, 0, 0, 1 a(n) = (-1)^n * A010815(n) q^(-1/24) * eta(q^2)^3 / (eta(q) * eta(q^4)) Euler transform of period 4: [[1,-2,1,-1] eps P="[2,3;1,-1;4,-1]"
Leading power of q factor (pqf)
Some of the generating functions have an additional power of q factor, in order to normalize the leading coefficient to q^0
.
- Example A226862:
Expansion of q^(-1/6) * eta(q^4) * eta(q^6)^5 / (eta(q^3) * eta(q^12))^2 eps P="[4,1;6,5;3,-2;12,-2]"
For all pairs we multiply the q power with the eta power and add the results, giving 4*1 + 6*5 - 3*2 - 12*2 = 4 + 30 - 6 - 24 = 4. This is the negative numerator of the pqf (expanded for denominator 24).
- Example A286813:
q^(-9/8) * (eta(q^2) * eta(q^16))^2 / (eta(q) * eta(q^8)) eps P="[2,2;16,2;1,-1;8,-1]" gives 2*2 * 16*2 - 1*1 - 8*1 = 36 - 9 = 27 and pqf = q^(-27/24)
- Example A286134 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, -1, -1, -1
q^(-1/2) * eta(q^5) * eta(q^6) * eta(q^7) * eta(q^210) eps P="[5,1;6,1;7,1;210,1]" gives 18 + 210 = 228 and pqf = q^(-228/24) = -9 -1/2.
The sequence is shifted right by 9 positions in this case.
Normalization of an EPS
For long lists of generating functions it is convenient to require all EPSs to be normalized, such that they can be tested for equality and can be sorted. We use the following rules:
- The sublist of pairs with positive ei comes before the sublist with negative ei.
- Inside a sublist the first components qi are in descending order.
- Successive pairs with the same qi are combined into a single pair with the sum of the ei as second component.
Exponentiation of an EPS
An EPS is taken to a power m by simply multiplying all qi (the first components in the pairs) by m. The exponent may be 1/2 for the square root, or any other fraction mn/md as long as the qi are all divisible by md.
Multiplication of two EPSs
Two EPSs could be multiplied by simply concatenating their list of pairs. It is convenient, howver, to combine all occurrences of specific qi by adding the corresponding eta exponents e.