An averaging formula for Nielsen numbers of affine n-valued maps on infra-nilmanifolds

This paper establishes an averaging formula to compute the Nielsen number of any $n$-valued affine map on an infra-nilmanifold, extending previous results known for single-valued maps and $n$-valued maps on nilmanifolds.

The Gist

Here is an explanation of the paper, translated from complex mathematical jargon into a story about maps, mirrors, and counting.

The Big Picture: Counting the Unavoidable

Imagine you have a magical map of a city (a manifold). You decide to play a game: you take every single person in the city and tell them to move to a new location.

In the world of Nielsen Theory, mathematicians are obsessed with one specific question: "No matter how you shuffle the people around, is there a minimum number of people who must end up standing exactly where they started?"

This minimum number is called the Nielsen Number. It's like a guarantee. Even if you try your hardest to move everyone away from their starting spot, the laws of geometry and topology say, "Sorry, at least X people have to stay put."

The Cast of Characters

1. The City (Infra-nilmanifold): Think of this as a weird, twisted city. It might look like a flat sheet of paper (a torus) in some places, but if you walk far enough, you might end up back where you started, but upside down or mirrored. The Klein Bottle mentioned in the paper is a famous example of such a city—it's a surface with no "inside" or "outside."
2. The Multi-Valued Map (The $n$-valued map): In the old days, a map told you to move from Point A to exactly one Point B. But in this paper, the map is a bit chaotic. It says, "From Point A, you can go to any of these $n$ different spots." It's like a "Choose Your Own Adventure" book where every page has $n$ possible next steps.
3. The Affine Map (The Algebraic Rule): The authors focus on a specific, orderly type of chaos. Instead of random jumps, the movement follows a strict algebraic recipe (like a robot following a formula). This makes the problem solvable.

The Problem: Why is this hard?

For simple, flat cities (called nilmanifolds, like a standard donut shape), mathematicians already had a magic formula to calculate the Nielsen Number. It was like having a calculator that just gave you the answer.

But for the twisted, mirrored cities (infra-nilmanifolds), the old formula broke.

• The Old Trick: Previously, to solve a problem on a twisted city, you would "unroll" it into a flat sheet (a covering space), solve the problem there, and then average the results.
• The New Problem: With these multi-valued maps (where one person splits into $n$ paths), you can't always "unroll" the map onto the flat sheet. The paths get tangled in a way that doesn't fit on the flat sheet. The old trick fails.

The Solution: The "Averaging Formula"

The authors (Karel and Lore) came up with a new, clever way to count. Instead of trying to unroll the whole map, they realized they could break the twisted city into small, manageable chunks.

The Analogy of the Mirror Maze:
Imagine the twisted city is a hall of mirrors.

1. The Reflections: The city has a "core" version (the flat sheet) that it covers multiple times. Think of the core as the "real" room, and the twisted city as a hall of mirrors reflecting that room.
2. The Average: The authors proved that to find the guaranteed number of fixed points in the twisted city, you don't need to look at the whole mess at once.
   • You look at the "core" room.
   • You look at every possible way the mirrors can reflect that room (the different "twists" or symmetries).
   • For each reflection, you calculate how many fixed points would exist if the map were simple.
   • The Magic: You take all those numbers, add them up, and divide by the number of reflections.

The Formula in Plain English:

"The number of unavoidable fixed points in the twisted city is the average of the fixed points you would find in all the different 'mirror versions' of the flat city."

The "Klein Bottle" Example

To prove their formula works, they tested it on the Klein Bottle (a twisted city with no inside/outside).

• They created a specific 2-valued map (a rule that sends every point to two possible destinations).
• They tried to use the old "unrolling" method, but it failed because the map didn't fit neatly onto the flat torus that covers the Klein bottle.
• They applied their new Averaging Formula.
• The Result: The formula predicted exactly 1 fixed point.
• The Verification: They did the math manually and found that, indeed, no matter how you look at it, there is exactly one spot where the rules force a person to stay put.

Why This Matters

This paper is a bridge. It connects the simple, flat world of mathematics (where we have easy formulas) to the complex, twisted world (where things get messy).

• Before: We could count fixed points on flat surfaces, but we were stuck when the surface was twisted and the map was multi-valued.
• Now: We have a universal calculator (the Averaging Formula) that works for any "affine" map on these twisted surfaces.

It's like discovering that even if you are walking through a confusing, mirrored maze, you can still predict exactly how many times you will bump into your own reflection, simply by averaging the results of the straight paths that make up the maze.

Technical Summary

Here is a detailed technical summary of the paper "An Averaging Formula for Nielsen Numbers of Affine n-Valued Maps on Infra-Nilmanifolds" by Karel Dekimpe and Lore De Weerdt.

1. Problem Statement

The paper addresses the computation of the Nielsen number $N(f)$, a homotopy invariant that provides a sharp lower bound for the number of fixed points of a map $f$ homotopic to a given map.

• Context: The study focuses on infra-nilmanifolds (quotients of simply connected nilpotent Lie groups by almost-Bieberbach groups) and $n$-valued maps (continuous set-valued functions mapping a point to a set of $n$ distinct points).
• Gap in Literature:
  • For single-valued maps on infra-nilmanifolds, an averaging formula exists (Kim, Lee, Lee [8, 9]) which expresses $N(f)$ as the average of Nielsen numbers of lifts to a finite covering nilmanifold.
  • For single-valued maps on nilmanifolds, the formula for $n$-valued maps was established in [4].
  • The Challenge: In the $n$-valued case, it is not generally true that any $n$-valued map on an infra-nilmanifold is homotopic to an affine $n$-valued map, nor do these maps necessarily admit lifts to a covering nilmanifold in the same way single-valued maps do. Consequently, the standard averaging approach via lifts fails.
• Objective: To establish a generalized averaging formula to compute the Nielsen number of any affine $n$-valued map on an infra-nilmanifold without relying on the existence of a global lift to a nilmanifold.

2. Methodology

The authors employ a combination of algebraic topology, group theory, and Lie algebra analysis.

A. Nielsen Theory for $n$-Valued Maps

The authors utilize the framework developed by Brown et al. [1], where an $n$-valued map $f: X \to D_n(X)$ is analyzed via its lifts to the orbit configuration space $F_n(\tilde{X}, \pi)$.

• A lift $\tilde{f}$ splits into $n$ single-valued maps $\tilde{f}_1, \dots, \tilde{f}_n$.
• These induce a morphism $(\phi_1, \dots, \phi_n; \sigma): \pi \to \pi^n \rtimes \Sigma_n$.
• The Nielsen number is initially expressed as a sum over Reidemeister classes $R[\phi_i]$ and $\sigma$-classes (Equation 1.1).

B. Decomposition of the Fixed Point Set

To handle the infra-nilmanifold structure $\pi \backslash G$ (where $G$ is a nilpotent Lie group and $\pi$ is an almost-Bieberbach group), the authors decompose the general averaging formula.

• They consider a finite covering nilmanifold $N \backslash G$, where $N = \pi \cap G$ is a finite index normal subgroup of $\pi$.
• They decompose the Reidemeister classes of the morphisms $\phi_i$ (defined on $\pi$) into classes associated with the quotient group $\pi/N$ and the subgroup $N$.
• Key Technical Step: Using Lemma 3.1 (from [7]), they relate the number of Reidemeister classes in the larger group $\pi$ to those in the subgroup $N$ and the finite quotient $\pi/N$. This involves analyzing coincidence groups and stabilizers of the action of $\pi/N$ on the quotient space.

C. Analysis of Affine Maps

The authors restrict their focus to affine $n$-valued maps, which lift to maps of the form:
$$\tilde{f}(x) = (g_1 \phi_1(x), \dots, g_n \phi_n(x))$$
where $g_i \in G$ and $\phi_i \in \text{End}(G)$.

• They analyze the induced Lie algebra morphisms $(\phi_i)_*$ and the linear parts $A_\alpha$ of the affine transformations in the group $\pi \subset G \rtimes \text{Aut}(G)$.
• Crucial Lemmas:
  • Lemma 4.1 & 4.2: Establish conditions under which fixed points exist or are unique based on the determinant of $(I - (\phi')_*)$.
  • Lemma 4.5: Proves that if the linear part has an eigenvalue of 1, the map can be homotoped to have no fixed points (index 0).
  • Lemma 4.7: Provides a formula for the Reidemeister number $R(\phi')$ in terms of the determinant of the Lie algebra morphism.
  • Lemma 4.8: Shows that the determinant of the twisted morphism $(\tau_{g\alpha}\phi'_i)_*$ is invariant and equal to $\det(I - (A_\alpha \phi_i)_*)$.

3. Key Contributions and Results

Main Result: The Averaging Formula (Theorem 4.9)

The paper derives the following formula for the Nielsen number of an affine $n$-valued map $f$ on an infra-nilmanifold $\pi \backslash G$:

$$N(f) = \frac{1}{[\pi : N]} \sum_{\bar{\alpha} \in \pi/N} \sum_{i=1}^n \left| \det(I - (A_\alpha \phi_i)_*) \right|$$

Where:

• $N$ is a finite index nilpotent subgroup of $\pi$ (defining the covering nilmanifold).
• $\pi/N$ is the finite quotient group.
• $\bar{\alpha}$ represents elements of the quotient.
• $A_\alpha$ is the linear part of the affine transformation corresponding to $\alpha$.
• $\phi_i$ are the endomorphisms defining the affine map.
• $(A_\alpha \phi_i)_*$ denotes the induced morphism on the Lie algebra $\mathfrak{g}$.

Significance of the Formula

1. Generalization: It naturally generalizes the single-valued averaging formula from [8, 9] and the nilmanifold $n$-valued formula from [4].
2. Algebraic Computation: It reduces the topological problem of counting fixed points to a purely algebraic computation involving determinants of matrices derived from the Lie algebra.
3. No Lift Required: Unlike the single-valued case, this formula does not require the $n$-valued map to actually lift to the covering nilmanifold. It works by averaging over the finite quotient $\pi/N$ using the algebraic data of the map's lift to the universal cover $G$.

4. Illustrative Example

The authors apply the formula to a 2-valued map on the Klein bottle (viewed as an infra-nilmanifold $\pi \backslash \mathbb{R}^2$).

• Setup: The map is defined by specific translations and linear projections.
• Calculation: They compute the sum over the two elements of $\pi/Z$ (where $Z$ is the torus subgroup). The calculation yields $N(f) = 1$.
• Verification: A direct geometric calculation confirms the map has exactly one fixed point class.
• Critical Observation (Remark 5.1): The authors demonstrate that this specific map does not lift to any finite covering torus (nilmanifold).
  • If a lift existed, the induced morphism on the fundamental group would map the subgroup $N$ into $N^2 \rtimes \Sigma_2$.
  • They show that for this specific map, the image of a specific element $b^2$ under the induced morphism violates this condition.
  • Conclusion: This proves that the classical "lift-and-average" strategy used for single-valued maps is impossible for $n$-valued maps, validating the necessity of the new algebraic decomposition method presented in the paper.

5. Significance

This paper resolves a significant gap in fixed point theory for multi-valued maps on infra-nilmanifolds. By establishing that the Nielsen number can be computed via an averaging formula over the finite quotient group using Lie algebra determinants, the authors provide a robust tool for analyzing fixed points in a setting where topological lifting is obstructed. This bridges the gap between the algebraic simplicity of nilmanifolds and the more complex topology of infra-nilmanifolds in the context of $n$-valued dynamics.