REVIEW
A note on the Zassenhaus product formula
Mathematical Physics · Submitted 6 Mar 2006 · 11 references
Pith's verdict: A clean, honest extension of Reinsch's matrix method to Zassenhaus exponents; the recursion is sound, the conjecture is clearly labeled, and the few soft spots are minor.
Why Pith says this
Highlight any text to comment on it.
Referee report
The paper
The paper claims
The paper gives a finite matrix recursion that computes every Zassenhaus exponent c_n in the Zassenhaus product formula.
The reading
What carries the argument
The load-bearing object is the triple of $(n+1)\times(n+1)$ upper triangular matrices $H=\exp(P+Q)$, $K=\exp(-P)$, $L=\exp(-Q)$, where P is the superdiagonal shift matrix and Q is the same shift weighted by $\tau_i$ on row i. Since P and Q are nilpotent, all exponentials terminate in finite sums. The argument's mechanics are: (i) write the Zassenhaus identity at matrix level; (ii) peel off $e^{C_2},\dots,e^{C_{n-1}}$ by multiplying by their inverses; (iii) read the surviving upper-right matrix element, which by the cited lemma is the polynomial in $\tau$ whose monomials describe length-n words; and (iv) apply U to turn each $\tau$-monomial into the corresponding word in a and b. The translation U is the bridge between the commuting auxiliary variables and the non-commuting letters of the original problem.
Extended reading notes
Core claim
The central claim is Corollary 1. For n>2, the Zassenhaus exponent c_n is obtained as $c_n = U\!\left((e^{-C_{n-1}} \cdots e^{-C_2} L K H)_{1,n+1}\right)$, where C_m denote the previously computed exponents written as matrices in P and Q, L, K, H are the three triangular matrices defined by $L=\exp(-Q)$, $K=\exp(-P)$, $H=\exp(P+Q)$, and U translates a monomial in the commuting $\tau_i$ into the word whose i-th letter is b or a according as $\tau_i$ appears or not. The same formula with n=2 uses just LKH. The argument proceeds by a short matrix calculation: from the exact identity $e^{P+Q}=e^P e^Q \prod_i e^{C_i}$ one isolates $e^{C_n}$ as $e^{-C_{n-1}} \cdots e^{-C_2} e^{-Q} e^{-P} e^{P+Q}$; because the matrices are nilpotent shifts, only the upper-right entry survives and records the word coefficients. The note also states a conjecture giving c_n as a sum over words starting with ba (or ab) of $c(W)/n_b(W)$ (or $c(W)/n_a(W)$) times a left-nested commutator, checked up to n=17.
Load-bearing premise
The recursion depends on the lemma, cited from the literature and not proved here, that the upper-right corner of any product of n shift matrices P and Q is a monomial in the tau_i whose nonzero exponents mark exactly the positions of the Q factors; only through that identification does U turn matrix entries into words in a and b.
Editorial extensions
If this is right
- Every Zassenhaus exponent c_n can be computed, in principle to arbitrary order, by finite matrix multiplication and the word-translation U, without solving the exponential identity order by order.
- The provided Mathematica implementation gives a concrete tool: one call returns c_n in words, and the authors report reproducing standard expressions and reaching n=17 on a 2 GB machine in about an hour.
- If the conjectured commutator representation is correct, Zassenhaus exponents possess substantially smaller nested-commutator descriptions than the generic word-to-commutator translation yields.
- The same triangular-matrix scheme is stated to extend to q-deformed exponentials, carrying the method beyond ordinary Lie-algebra words.
Reading between the lines
- An extension left implicit: the same peeling procedure should yield Zassenhaus exponents for more than two non-commuting factors, since the matrix identity $e^{P+Q}=e^P e^Q \prod_i e^{C_i}$ can be generalized to a product of several exponentials.
- A testable extension: because U is linear and invertible, the matrix product effectively encodes the full distribution of word coefficients, so one could read individual coefficients without expanding the entire c_n.
- If the conjecture (16) holds for all n, the two equivalent sums over ba-leading and ab-leading words reflect an antisymmetry that could be used to halve storage in high-order symbolic computation.
- The recurrence can be viewed as a triangular factorization of $\exp(P+Q)$; this viewpoint suggests a direct algebraic proof of the compact commutator form by showing the two sums differ only by Jacobi identities.
Original abstract
We provide a simple method for the calculation of the terms c_n in the Zassenhaus product $e^{a+b}=e^a e^b \prod_{n=2}^{\infty} e^{c_n}$ for non-commuting a and b. This method has been implemented in a computer program. Furthermore, we formulate a conjecture on how to translate these results into nested commutators. This conjecture was checked up to order n=17 using a computer.
The compiled record
The paper's own sentences, arranged, kept beside the review and separate from it. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
The abstract, sentence by sentence
We provide a simple method for the calculation of the terms c_n in the Zassenhaus product $e^{a+b}=e^a e^b \prod_{n=2}^{\infty} e^{c_n}$ for non-commuting a and b. This method has been implemented in a computer program. Furthermore, we formulate a conjecture on how to translate these results into nested commutators. This conjecture was checked up to order n=17 using a computer.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | We provide a simple method for the calculation of the terms c_n in the Zassenhaus product $e^{a+b}=e^a e^b \prod_{n=2}^{\infty} e^{c_n}$ for non-commuting a and b. | contribution 0.97 | 0.94 | 0.66 | asserted 1.00 |
| s1 | This method has been implemented in a computer program. | method 0.73 | 0.90 | 0.05 | asserted 1.00 |
| s2 | Furthermore, we formulate a conjecture on how to translate these results into nested commutators. | contribution 0.85 | 0.91 | 0.14 | speculative 0.92 |
| s3 | This conjecture was checked up to order n=17 using a computer. | result 0.47 | 0.89 | 0.10 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
References
Discussion
Disagree with the review, or know something it missed? Leave a comment, highlight any passage above to comment on it, or challenge this review formally.