[Paper Review] How the permutation of edges of a metric graph affects the number of points moving along the edges

В. Л. Чернышев, А. А. Толченников|arXiv (Cornell University)|Oct 18, 2014

Graph Labeling and Dimension Problems8 references3 citations

TL;DR

This paper investigates how edge permutation in a metric graph affects the number of points moving along its edges, modeling a semiclassical Schrödinger dynamics via a discrete system. By reducing the problem to counting non-negative integer solutions of weak inequalities, it derives explicit asymptotic expansions for the number of moving points, showing that edge permutation induces a $ T^3 $-order difference in the count, with the leading coefficient depending on propagation times of all edges except a 'jumper' edge.

ABSTRACT

We consider a dynamical system on a metric graph, that corresponds to a semiclassical solution of a time-dependent Schrödinger equation. We omit all details concerning mathematical physics and work with a purely discrete problem. We find a weak inequality representation for the number of points coming out of the vertex of an arbitrary tree graph. We apply this construction to an "H-junction" graph. We calculate the difference between numbers of moving points corresponding to the permutation of edges. Then we find a symmetrical difference of the number of points moving along the edges of a metric graph.

Motivation & Objective

• To analyze the asymptotic behavior of moving point counts on metric graphs under time evolution.
• To understand how topological rearrangements (edge permutations) affect the number of points moving along edges.
• To derive explicit asymptotic expansions for the number of points, focusing on the second-order term in the expansion.
• To introduce and compute a symmetric difference in point counts over all internal vertex pairs to eliminate initial condition bias.
• To validate analytical results through direct computation and computer experiments with incommensurate edge times.

Proposed method

• Model the dynamics on a tree graph as a discrete system where points split at vertices based on edge valence and propagate with fixed times.
• Express the number of points arriving by time $ T $ as the number of non-negative integer solutions to weak inequalities derived from closed paths returning to the root.
• Use geometric interpretation of these solutions as lattice points in expanding simplices of dimension four.
• Apply known asymptotic expansions for lattice point counts in simplices, incorporating terms from the Sprague-Grundy function and number-theoretic functions.
• Derive the leading and second-order terms in the asymptotic expansion of the number of points, using symmetric sums over all internal vertices to eliminate dependence on initial vertex choice.
• Validate analytical results via direct computation and numerical experiments with incommensurate edge times $ t_i = 1, √2, √3, √5, √7 $.

Experimental results

Research questions

• RQ1 How does permuting the edges of a metric graph affect the number of points moving along its edges over time?
• RQ2 What determines the second-order term in the asymptotic expansion of the number of moving points on a tree graph?
• RQ3 Can the symmetric difference of point counts over all internal vertex pairs eliminate dependence on the initial vertex?
• RQ4 What is the leading-order coefficient of the difference in point counts between two graphs with identical edge sets but different edge permutations?
• RQ5 How do the propagation times of edges influence the asymptotic difference in point counts, particularly in the $ T^3 $-order term?

Key findings

• The difference in the number of moving points between two graphs differing only by edge permutation is of order $ T^3 $, with the leading coefficient explicitly expressed in terms of propagation times of all edges except a 'jumper' edge.
• For an H-junction graph, the difference $ N( ilde{Γ}_H, A, T) - N( ilde{Γ}'_H, A, T) = -\frac{1}{96 t_2 t_4} \left( \frac{1}{t_5} - \frac{1}{t_1} \right) T^3 + o(T^3) $, showing explicit dependence on four edge times.
• The symmetric difference over all pairs of internal vertices is of order $ T^2 $, with leading coefficient $ -\frac{1}{4} \left( \frac{1}{t_1 t_2} + \frac{1}{t_4 t_5} - \frac{1}{t_2 t_4} - \frac{1}{t_1 t_5} \right) T^2 + o(T^2) $, eliminating initial condition bias.
• The second-order term in the asymptotic expansion depends on the initial vertex, which justifies the use of symmetric difference to isolate intrinsic graph properties.
• Analytical results were confirmed by direct computation and computer experiments using incommensurate edge times $ t_1=1, t_2=\sqrt{2}, t_3=\sqrt{3}, t_4=\sqrt{5}, t_5=\sqrt{7} $, showing full agreement with theory.
• The assumption of linear independence over $ \mathbb{Q} $ ensures the validity of the asymptotic expansions, and numerical evidence supports its applicability for almost all such edge sets.