Overview
This study examines the spectral properties of trees, specifically focusing on combinations of their largest ($\lambda_1(G)$) and second largest ($\lambda_2(G)$) adjacency eigenvalues. The investigation covers the sum of these eigenvalues, known as the spectral sum ($\lambda_1(G) + \lambda_2(G)$), and a more general convex combination ($\alpha \lambda_1 + (1-\alpha)\lambda_2$), where $\alpha \in [0,1]$. The primary objective is to identify the structural characteristics of trees that yield maximum or minimum values for these spectral measures.
Research Context
In graph theory, eigenvalues of adjacency matrices provide insights into the structure and properties of graphs. For a graph $G$, $\lambda_1(G)$ represents the largest adjacency eigenvalue, and $\lambda_2(G)$ denotes the second largest adjacency eigenvalue. These values are fundamental in spectral graph theory. The study specifically targets trees, which are connected graphs without cycles, and considers their order $n$, referring to the number of vertices. The concept of a 'spectral sum' or a 'convex combination' of eigenvalues extends the analysis beyond individual eigenvalues, exploring how their combined magnitude varies across different tree structures.
Approach
The research employed an analytical approach to investigate the spectral sum $\lambda_1(G) + \lambda_2(G)$ and the convex combination $\alpha \lambda_1 + (1-\alpha)\lambda_2$ within the class of trees of order $n$. The methodology involved:
- Determining the extremal trees that achieve the maximum value for $\lambda_1(G) + \lambda_2(G)$ among all trees of order $n$.
- Identifying the extremal trees that attain the minimum value for $\lambda_1(G) + \lambda_2(G)$ among all trees of order $n$.
- Characterizing the extremal trees that maximize the convex combination $\alpha \lambda_1 + (1-\alpha)\lambda_2$ for any given $\alpha \in [0,1]$, specifically for cases where $n$ is sufficiently large.
Findings
- The study determined the extremal trees for the spectral sum $\lambda_1(G) + \lambda_2(G)$ within the class of trees of order $n$. These findings specify the tree structures that exhibit the highest and lowest spectral sums.
- For any $\alpha \in [0,1]$, the research described the extremal trees that maximize the convex combination $\alpha \lambda_1 + (1-\alpha)\lambda_2$. This characterization is valid for $n$-vertex trees when $n$ is sufficiently large.