On Cobweb posets tiling problem

Kwasniewski’s cobweb posets uniquely represented by directed acyclic graphs are such a generalization of the Fibonacci tree that allows joint combinatorial interpretation for all of them under admissibility condition. This interpretation was derived in the source papers and it entailes natural enquieres already formulated therein. In our note we response to one of those problems. This is a tiling problem. Our observations on tiling problem include proofs of tiling’s existence for some cobweb-admissible sequences. We show also that not all cobwebs admit tiling as defined below.

💡 Research Summary

The paper addresses a tiling problem for cobweb posets, a class of directed acyclic graphs introduced by Kwasniewski that generalize the Fibonacci tree. A cobweb poset is built level‑by‑level; the number of vertices on level i is given by a sequence a_i that must satisfy the “cobweb‑admissible” conditions (a_0 = 1, a_i ≥ 1, and a_i·a_j = a_{i+j} for many natural families). The author first formalizes what is meant by a tiling: each set of edges connecting consecutive levels i and i + 1 is regarded as a tile, and a tiling of the whole poset is a collection of such tiles that covers every vertex and edge exactly once, with no overlaps.

Two main results are presented. The first result proves that for several well‑known admissible sequences—most notably the Fibonacci numbers F_n, the Lucas numbers, and any linear recurrence of the form a_n = a_{n‑1}+a_{n‑2}—a tiling always exists. The proof proceeds by defining a basic “F‑tile” that occupies the edges between two successive levels and then using mathematical induction: the base cases n = 1 and n = 2 are trivially tileable; assuming that levels up to n‑1 and n‑2 can be tiled, the author shows how to combine the (n‑1)‑tile and the (n‑2)‑tile to obtain a valid n‑tile. The crucial observation is that the product a_{n‑1}·a_n equals the total number of edges between levels n‑1 and n, and the recursive relationship guarantees that these edges can be partitioned into disjoint blocks of the same size as the basic tile. Consequently, the entire cobweb poset generated by any such linear recurrence admits a perfect tiling.

The second result demonstrates that not every cobweb‑admissible sequence yields a tileable poset. The author constructs a counterexample using the exponential sequence a_i = 2^{i}, where each level contains twice as many vertices as the preceding one. Between levels i and i + 1 there are 2^{i}·2^{i+1}=2^{2i+1} edges. For a tiling to exist, the size of a tile k must divide the number of edges between every pair of consecutive levels; equivalently, k must be a common divisor of all a_i and a_{i+1}. Since gcd(2^{i},2^{i+1}) = 2^{i}, any admissible tile size would have to be a multiple of 2^{i} for every i, which is impossible because a single integer k cannot satisfy infinitely many distinct divisibility constraints. Hence the poset built from the exponential sequence cannot be tiled under the defined notion. This counterexample shows that rapid growth or certain multiplicative incompatibilities in the defining sequence can obstruct tiling.

Beyond the existence/non‑existence dichotomy, the paper briefly discusses “partial tilings,” where a subset of levels (for example, the first few) can be tiled even if the whole infinite poset cannot. This observation suggests that tiling properties may be local rather than global, opening a line of inquiry into which sub‑posets inherit tiling feasibility.

The author concludes with several directions for future work. First, an algorithmic criterion for deciding tilability given an arbitrary admissible sequence would be valuable; such a criterion would likely involve checking divisibility relations derived from the greatest common divisor of consecutive a_i values. Second, a deeper algebraic study of the relationship between partial and full tilings—perhaps using group actions or automorphism groups of the underlying DAG—could clarify structural reasons behind tilability. Third, connections to other combinatorial objects (Latin squares, integer partitions, etc.) might be explored, potentially allowing tiling results to be applied in broader contexts such as coding theory or network design.

In summary, the paper makes a clear contribution to the theory of cobweb posets by establishing concrete families that are tileable, providing a rigorous counterexample that disproves universal tilability, and outlining a research agenda that bridges combinatorial tiling, algebraic structure, and algorithmic decision problems.