Witt symmetric functions¶
-
sage.combinat.sf.witt.
SymmetricFunctionAlgebra_witt
¶ The Witt symmetric function basis (or Witt basis, to be short).
The Witt basis of the ring of symmetric functions is denoted by \((x_{\lambda})\) in [HazWitt1], section 9.63, and by \((q_{\lambda})\) in [DoranIV1996]. We will denote this basis by \((w_{\lambda})\) (which is precisely how it is denoted in [GriRei18], Exercise 2.9.3(d)). It is a multiplicative basis (meaning that \(w_{\emptyset} = 1\) and that every partition \(\lambda\) satisfies \(w_{\lambda} = w_{\lambda_1} w_{\lambda_2} w_{\lambda_3} \cdots\), where \(w_i\) means \(w_{(i)}\) for every nonnegative integer \(i\)).
This basis can be defined in various ways. Probably the most well-known one is using the equation
\[\prod_{d=1}^{\infty} (1 - w_d t^d)^{-1} = \sum_{n=0}^{\infty} h_n t^n\]where \(t\) is a formal variable and \(h_n\) are the complete homogeneous symmetric functions, extended to \(0\) by \(h_0 = 1\). This equation allows one to uniquely determine the functions \(w_1, w_2, w_3, \ldots\) by recursion; one consequently extends the definition to all \(w_{\lambda}\) by requiring multiplicativity.
A way to rewrite the above equation without power series is:
\[h_n = \sum_{\lambda \vdash n} w_{\lambda}\]for all nonnegative integers \(n\), where \(\lambda \vdash n\) means that \(\lambda\) is a partition of \(n\).
A similar equation (which is easily seen to be equivalent to the former) is
\[e_n = \sum_{\lambda} (-1)^{n - \ell(\lambda)} w_{\lambda},\]with the sum running only over strict partitions \(\lambda\) of \(n\) this time. This equation can also be used to recursively define the \(w_n\). Furthermore, every positive integer \(n\) satisfies
\[p_n = \sum_{d\mid n} d w_d^{n/d},\]and this can be used to define the \(w_n\) recursively over any ring which is torsion-free as a \(\ZZ\)-module. While these equations all yield easy formulas for classical bases of the ring of symmetric functions in terms of the Witt symmetric functions, it seems difficult to obtain explicit formulas in the other direction.
The Witt symmetric functions owe their name to the fact that the ring of symmetric functions can be viewed as the coordinate ring of the group scheme of Witt vectors, and the Witt symmetric functions are the functions that send a Witt vector to its components (whereas the powersum symmetric functions send a Witt vector to its ghost components). Details can be found in [HazWitt1] or section 3.2 of [BorWi2004].
INPUT:
Sym
– an instance of the ring of the symmetric functions.coerce_h
– (default:True
) a boolean that determines whether the transition maps between the Witt basis and the complete homogeneous basis will be cached and registered as coercions.coerce_e
– (default:False
) a boolean that determines whether the transition maps between the Witt basis and the elementary symmetric basis will be cached and registered as coercions.coerce_p
– (default:False
) a boolean that determines whether the transition maps between the Witt basis and the powersum basis will be cached and registered as coercions (or conversions, if the base ring is not a \(\QQ\)-algebra).
REFERENCES:
[HazWitt1] (1, 2) Michiel Hazewinkel. Witt vectors. Part 1. Arxiv 0804.3888v1 [DoranIV1996] William F. Doran IV. A Proof of Reutenauer’s `-q_{(n)}` Conjecture. Journal of combinatorial theory, Series A 74, pp. 342-344 (1996), article no. 0056. doi:10.1006/jcta.1996.0056 [BorWi2004] James Borger, Ben Wieland. Plethystic algebra. Arxiv math/0407227v1 EXAMPLES:
Here are the first few Witt symmetric functions, in various bases:
sage: Sym = SymmetricFunctions(QQ) sage: w = Sym.w() sage: e = Sym.e() sage: h = Sym.h() sage: p = Sym.p() sage: s = Sym.s() sage: m = Sym.m() sage: p(w([1])) p[1] sage: m(w([1])) m[1] sage: e(w([1])) e[1] sage: h(w([1])) h[1] sage: s(w([1])) s[1] sage: p(w([2])) -1/2*p[1, 1] + 1/2*p[2] sage: m(w([2])) -m[1, 1] sage: e(w([2])) -e[2] sage: h(w([2])) -h[1, 1] + h[2] sage: s(w([2])) -s[1, 1] sage: p(w([3])) -1/3*p[1, 1, 1] + 1/3*p[3] sage: m(w([3])) -2*m[1, 1, 1] - m[2, 1] sage: e(w([3])) -e[2, 1] + e[3] sage: h(w([3])) -h[2, 1] + h[3] sage: s(w([3])) -s[2, 1] sage: Sym = SymmetricFunctions(ZZ) sage: w = Sym.w() sage: e = Sym.e() sage: h = Sym.h() sage: s = Sym.s() sage: m = Sym.m() sage: p = Sym.p() sage: m(w([4])) -9*m[1, 1, 1, 1] - 4*m[2, 1, 1] - 2*m[2, 2] - m[3, 1] sage: e(w([4])) -e[2, 1, 1] + e[3, 1] - e[4] sage: h(w([4])) -h[1, 1, 1, 1] + 2*h[2, 1, 1] - h[2, 2] - h[3, 1] + h[4] sage: s(w([4])) -s[1, 1, 1, 1] - s[2, 1, 1] - s[2, 2] - s[3, 1]
Some examples of conversions the other way:
sage: w(h[3]) w[1, 1, 1] + w[2, 1] + w[3] sage: w(e[3]) -w[2, 1] + w[3] sage: w(m[2,1]) 2*w[2, 1] - 3*w[3] sage: w(p[3]) w[1, 1, 1] + 3*w[3]
Antipodes:
sage: w([1]).antipode() -w[1] sage: w([2]).antipode() -w[1, 1] - w[2]
The following holds for all odd \(i\) and is easily proven by induction:
sage: all( w([i]).antipode() == -w([i]) for i in range(1, 10, 2) ) True
The Witt basis does not allow for simple expressions for comultiplication and antipode in general (this is related to the fact that the sum of two Witt vectors isn’t easily described in terms of the components). Therefore, most computations with Witt symmetric functions, as well as conversions and coercions, pass through the complete homogeneous symmetric functions by default. However, one can also use the elementary symmetric functions instead, or (if the base ring is a \(\QQ\)-algebra) the powersum symmetric functions. This is what the optional keyword variables
coerce_e
,coerce_h
andcoerce_p
are for. These variables do not affect the results of the (non-underscored) methods ofself
, but they affect the speed of the computations (the more of these variables are set toTrue
, the faster these are) and the size of the cache (the more of these variables are set toTrue
, the bigger the cache). Let us check that the results are the same no matter to what the variables are set:sage: Sym = SymmetricFunctions(QQ) sage: p = Sym.p() sage: wh = Sym.w() sage: we = Sym.w(coerce_h=False, coerce_e=True) sage: wp = Sym.w(coerce_h=False, coerce_p=True) sage: all( p(wh(lam)) == p(we(lam)) == p(wp(lam)) for lam in Partitions(4) ) True sage: all ( wh(p(lam)).monomial_coefficients() ....: == we(p(lam)).monomial_coefficients() ....: == wp(p(lam)).monomial_coefficients() for lam in Partitions(4) ) True