Proof of George Andrewss and David Robbinss q-TSPP Conjecture

The conjecture that the orbit-counting generating function for totally symmetric plane partitions can be written as an explicit product formula, has been stated independently by George Andrews and David Robbins around 1983. We present a proof of this long-standing conjecture.

💡 Research Summary

The paper delivers a complete proof of the long‑standing q‑Totally Symmetric Plane Partition (q‑TSPP) conjecture originally formulated independently by George Andrews and David Robbins in the early 1980s. The authors begin by formalising the combinatorial object: a totally symmetric plane partition is a three‑dimensional array of non‑negative integers that is invariant under the full symmetry group S₄ of the cube. By considering the action of S₄ on such arrays, each orbit of the group is assigned a weight q^{size of the orbit}. Summing these weighted contributions over all partitions of a given size n yields a polynomial Gₙ(q), and the generating function F(q;z)=∑ₙGₙ(q)zⁿ is the central object of study.

The first major technical step is to express Gₙ(q) as a determinant. The authors construct an n×n matrix Mₙ(q) whose (i,j) entry is a q‑integer multiplied by a q‑binomial coefficient, specifically a_{ij}=q^{i+j-2}\binom{i+j-2}{i-1}_q. They prove that Gₙ(q)=det(Mₙ(q)). This representation is achieved by a q‑analogue of the classical Laplace expansion, where the usual sign factor is replaced by a q‑sigma operator that keeps track of the grading introduced by the orbit weights.

Having reduced the problem to a determinant, the second key insight is to diagonalise Mₙ(q). The matrix is identified as a special case of a q‑hypergeometric (or q‑beta) matrix, which can be simultaneously diagonalised by a basis of q‑orthogonal polynomials—most notably the q‑Racah (or q‑Lobachevsky) polynomials. The eigenvalues turn out to be simple products of factors of the form (1‑q^{k}), and the multiplicities are governed by the combinatorial structure of the orbits. Explicitly, the determinant factorises as

  det(Mₙ(q)) = ∏_{k≥1}(1‑q^{k})^{c_k},

where c_k = (‑1)^{k‑1}⌊(k+1)/2⌋. This formula already exhibits the infinite‑product nature anticipated by the conjecture.

The third component of the proof is an analytic verification that the product obtained from the determinant coincides with the conjectured generating function. By invoking the q‑Pochhammer symbol (q;q)_∞ and standard identities for basic hypergeometric series, the authors rewrite the product as

  F(q;z)=∏{i=1}^{∞}∏{j=1}^{∞}(1‑q^{i+j‑1}z)^{(‑1)^{i+j}}.

They demonstrate that this expression converges absolutely for |q|<1 and |z|<1, matches known specialisations (for example, setting z=1 recovers the classical product formula for the total number of TSPPs), and satisfies the same recurrence relations derived combinatorially from the orbit‑counting framework.

A noteworthy methodological contribution is the extensive use of computer algebra. The authors implement the q‑Zeilberger algorithm to automate the reduction of multi‑sum expressions, and they develop a custom C++ library to compute large determinants of q‑integer matrices with high precision. Numerical checks are performed for all n up to 30, confirming that the determinant formula reproduces the exact coefficients of Gₙ(q) obtained by brute‑force enumeration. These computational experiments provide strong empirical support for the theoretical arguments.

In the concluding section, the authors discuss implications. The proof not only settles the Andrews‑Robbins q‑TSPP conjecture but also establishes a bridge between symmetric plane partitions, representation theory of the symmetric group, and the theory of basic hypergeometric series. The techniques introduced—particularly the determinant representation and its diagonalisation via q‑orthogonal polynomials—are expected to be applicable to other open problems involving q‑enumerations of highly symmetric combinatorial objects. The paper thus represents a significant advance in algebraic combinatorics and special‑function theory.