A sharp upper bound for the number of connected sets in any grid graph
Hongxia Ma, Xian'an Jin, Weiling Yang, Meiqiao Zhang
TL;DR
The paper develops a general framework to count connected vertex subsets in grid-like graphs. It first provides a transfer-matrix formula for $N(K_m\times P_n)$, enabling exact enumeration via $N(K_m\times P_n)=\sum_{k=1}^n (n-k+1)\cdot N_{m,k}$ with $N_{m,k}=\mathbf{u}\cdot T^{k-1}\mathbf{1}$. It then introduces a multistep-recurrence-based lower bound $N'_{m,n}$ for the number of connected sets in $K_m\times P_n$ that also hit every column, and proves a sharp upper bound $N(P_m\times P_n)\le \sum_{k=1}^n (N_{m,k}-N'_{m,k})(n-k+1)$, with equality for $m=3,4$. The sections 4.1 and 4.2 supply explicit, recursive schemes to compute $N(P_3\times P_n)$ and $N(P_4\times P_n)$, yielding exact counts for these special grid widths and illustrating the method's potential for larger widths. Overall, the work advances a structured, recurrence-based path toward Vince's question on formulas for grid-graph connected-set counts and offers practical enumeration for particular cases.
Abstract
A connected set in a graph is a subset of vertices whose induced subgraph is connected. Although counting the number of connected sets in a graph is generally a \#P-complete problem, it remains an active area of research. In 2020, Vince posed the problem of finding a formula for the number of connected sets in the $(n\times n)$-grid graph. In this paper, we establish a sharp upper bound for the number of connected sets in any grid graph by using multistep recurrence formulas, which further derives enumeration formulas for the numbers of connected sets in $(3\times n)$- and $(4\times n)$-grid graphs, thus solving a special case of the general problem posed by Vince. In the process, we also determine the number of connected sets of $K_{m}\times P_{n}$ by employing the transfer matrix method, where $K_{m}\times P_{n}$ is the Cartesian product of the complete graph of order $m$ and the path of order $n$.
