New Lower Bounds for Tverberg Partitions with Tolerance in the Plane
Abstract
Let be a set points in a -dimensional space. Tverberg’s theorem says that, if is at least , then can be partitioned into sets whose convex hulls intersect. Partitions with this property are called Tverberg partitions. A partition has tolerance if the partition remains a Tverberg partition after removal of any set of points from . Tolerant Tverberg partitions exist in any dimension provided that is sufficiently large. Let be the smallest value of such that tolerant Tverberg partitions exist for any set of points in . Only few exact values of are known.
In this paper we establish a new tight bound for . We also prove many new lower bounds on for and .
1 Introduction
The classical Tverberg’s theorem [18] states that a sufficiently large set of points in can be partitioned into subsets such that their convex hulls intersect. For the history and recent advances around Tverberg’s theorem, we refer the reader to [3, 4, 6, 13].
Theorem 1 (Tverberg [18]).
Any set of at least points in can be partitioned into sets whose convex hulls intersect, i.e.
(1) |
In 1972, David Larman proved the first tolerant version of Tverberg’s theorem: any set of points in can be partitioned into two sets and such that their convex hulls intersect with tolerance one, i.e., for any point , . García-Colín [8] generalized it for any tolerance. Soberón and Strausz [17] generalized it further for partitions into many sets. The general problem can be stated as follows.
Problem (Tverberg partitions with tolerance). Given positive integers , find the smallest positive integer such that any set of points in can be partitioned into subsets such that for any set of at most points
(2) |
We call a partition of into subsets -tolerant if condition (2) holds for any set of at most points. Several upper bounds for are known [10, 12, 14, 17]. Some lower bounds for are known [9, 2, 15].
Theorem 2 (Ramírez-Alfonsín [2]).
For any
(3) |
Theorem 3 (García-Colín and Larman [9]).
(4) |
Theorem 4 (Soberón [15]).
(5) |
In contrast to Tverberg’s theorem, most bounds are not known to be tight. The only known tight bounds are the following. Larman [12] proved for . Forge, Las Vergnas and Schuchert [7] proved . For all and dimension one, by Mulzer and Stein [14]. In this paper we establish a new tight bound for .
Theorem 5.
.
We also improve the bound (5) for the plane.
Theorem 6.
Let be a positive integer. For any integers and
(6) |
Remark. The bound of Theorem 6 can be stated without parameter
(7) |
If and then Equation (6) provides a lower bound . We further improve the bound for and .
Theorem 7.
For any , .
Several upper bounds for are known [8, 10, 16]. For example, Soberón [16] proved for fixed and . Therefore, it is interesting to find lower and upper bounds for . Theorems 5,6, and 7 provide new lower bounds for for .
Remark. Recently, we improved some lower bounds using computer pograms [5]. For example, for .
2 Preliminaries
We start with a simple observation.
Observation 8.
Every part in a -tolerant partition has size at least .
The following lemma is simple, yet very useful in proving that a partition is not -tolerant.
Lemma 9.
Let be a partition of a finite set in and let be a subset of a set such that
(8) |
and . Then the partition of is not -tolerant.
Proof.
Let . Then . Therefore
(9) |
by Equation 8 since one of the sets in the intersection is and every other set is a subset of . ∎
Two special cases of Lemma 9.
Lemma 10.
A partition of a finite set in is not -tolerant if one of the following conditions holds.
-
(1)
A set contains a vertex of and .
-
(2)
A set contains two points such that is an edge of and .
3 Proofs of Theorems 6 and 7
First, we prove Theorem 6.
Proof of Theorem 6.
Construct a set of points in the plane as follows. Let be a regular -gon with center at the origin . Place points at the vertices of and points close to the origin. More specifically, we assume that where is the distance from the origin to the line . Let . We assume that the vertices of are sorted in clockwise order, see Figure 1.

We show that any partition of into is not -tolerant. By Observation 8, we assume for all . If for some , then contains a point since . The partition is not -tolerant by Lemma 10(1).
It remains to consider the case where all sets in the partition have size at least . At most sets in the partition may have size . Since there exists a set of size such that . Since , there exist two points and with such that the vertices between and are in (we assume here that if ). There is a set such that none of these vertices are in . Therefore the partition is not -tolerant since . Note that the distance from the origin to any point in is smaller than the distance from the origin to the line . ∎
Next we prove Theorem 7.

Proof of Theorem 7.
Construct a set of points in the plane as follows. First, place points in convex position where points are in clockwise order. Then place three points inside the convex hull of such that point is close to edge (it is required that is inside triangles and ), see Figure 2. We show that any partition of into is not -tolerant.
Without loss of generality, we assume . Therefore, . By Observation 8, we assume . Suppose if . Since , contains a point of . The partition is not -tolerant by Lemma 10(1).
It remains to consider the case where . If set does not contain a point , then and . Then contains two consecutive points of . The partition is not -tolerant by Lemma 10(2). Therefore .
Suppose for some . If or then the partition is not -tolerant using or , respectively, and Lemma 9. Therefore, we assume that both points and are in .
Let . Then . By the previous argument, set contains at least pairs of consecutive points of (the pair for each point ). Note that and . We will show that set contains a pair of consecutive points of . Then, the partition is not -tolerant by Lemma 10(2).
If , then and set contains a pair of consecutive points of . Then set contains a pair of consecutive points of .
Suppose , say where . Then set contains , and . There are two intervals and in containing points of . Note that each interval contains an even number of points. If set does not contain a pair of consecutive points of , then set contains at most half of points in each interval. Then . This contradiction implies that set contains a pair of consecutive points of .
Finally, if , then there exists only one interval and set must contain a pair of consecutive points of ; otherwise . The theorem follows. ∎
4 Proof of Theorem 5
We found a special configuration of nine points to prove a lower bound for Theorem 5. For this, we checked point configurations of all order types for . Order types, introduced by Goodman and Pollack [11], are useful for characterizing the combinatorial properties of point configurations. An order type of a set of points in general position in the plane is defined using a mapping that assigns each triple of integers with the orientation (either clockwise or counter-clockwise) of the triple . Two point sets and have the same order type if there is a bijection from to preserving this map, i.e., for any three distinct points in , the orientation of and the orientation of are the same. Aichholzer et al. [1] established that there are 158,817 order types for . We developed a program for testing each of these point sets whether it admits a 2-tolerant partition into two sets. The proof of the tolerance of a configuration found by the program turns out to be not simple even using the tools from Section 2.

Lemma 11.
There exists a set of nine points in the plane that does not admit a 2-tolerant partition into two sets.

Proof.
Consider a set of nine points shown in Fig.3 where and . We will prove that any partition of is not 2-tolerant, i.e. there exists a set such that
We can assume that . By Observation 8, we assume . Suppose that . If , then take . If , then the partition of is not 2-tolerant by Lemma 10(1).
It remains to analyze partitions with . By Lemma 10(2), we assume that does not contain two consecutive points (assuming ). For example, this implies that contains at least one point of . Consider the following cases depending on .
Case 1. Set contains exactly one point of , say . Then . One of the sets , , , say set , does not contain . The partition is not 2-tolerant by Lemma 9 using .

Case 2. Set contains exactly two points of , i.e. for some and .
Case 2a. Suppose . By our assumption . We can assume that using and Lemma 9. So, .
We also can assume that using and Lemma 9. Therefore . If (so ) then take , see Fig.4(a). If (so ) then take , see Fig.4(b).

Case 2b. Suppose . Then . We can assume using and Lemma 9. We can assume using and Lemma 9. Thus, . We can assume using and Lemma 9.
Case 2c. Suppose . We can assume using for and Lemma 9. Since and cannot be in together, there are three options. If , take , see Fig.5(b). If then we can use and Lemma 9. If then we can use and Lemma 9.
Case 3. Set contains exactly three points of , i.e. or .

Corollary 12.
.
Soberón and Strausz [17] proved an upper bound which implies . Therefore .
5 Further Discussion and Open Problems
Improving the bounds for Tverberg partitions with tolerance is an interesting problem. There are many triples with a gap between known lower and upper bounds. Finding sharp bounds for is a challenging problem. For example, Larman [12] conjectured that for any . The conjecture is open for . For dimension , the best known upper bound is by Larman [12] and the best known lower bound is by García-Colín and Larman [9].
Another possibility for a new sharp bound is for and tolerance . The lower bound is by Theorem 7. The best upper bound is by Soberón and Strausz [17]. Based on our computer experiments, we conjecture that . One way to improve the upper bound for and is to study the following core lemma in the proof of the upper bound for [17].
Lemma 13.
Let and be integers, be a set of points and be a group with . If there is an action of in a set which is compatible with , then for each there is a such that the set captures the origin with tolerance .
For and tolerance 3, Lemma 13 uses points since . Is this bound for sharp?
5.1 Order types and tolerant Tverberg partitions
Our proof of Theorem 5 uses the set of nine points shown in Figure 3. It was computed by a program that we developed for testing the tolerance of a given point set. We have applied this program to 158,817 order types for which are provided by Aichholzer et al. [1]. Our program found that 155,115 point sets admit a 2-tolerant partition into two sets. Only 3,702 point sets do not admit a 2-tolerant partition into two sets and therefore, can be used as examples for proving the lower bound .
We analyzed these 3,702 order types and found that their convex hulls have sizes between three and six. We count the number of different order types for each size of the convex hull, see Table 1. Figures 8, 9, and 10 show some order types with 5, 4, and 3 points on the convex hull, respectively. It follows from our computer experiments that any set of nine points in the plane with at least seven points on the convex hull admits a 2-tolerant partition into two sets. It is interesting to investigate how the size of convex hull of a point set affects lower bounds of .
1303 | 1554 | 769 | 76 | 0 | 0 | 0 |
References
- [1] O. Aichholzer, F. Aurenhammer, and H. Krasser. Enumerating order types for small point sets with applications. Order, 19:265–281, 2002.
- [2] J. L. R. Alfonsín. Lawrence oriented matroids and a problem of mcmullen on projective equivalences of polytopes. Eur. J. Comb., 22(5):723–731, 2001.
- [3] I. Bárány, P. V. M. Blagojevic, and G. M. Ziegler. Tverberg’s theorem at 50: extensions and counterexamples. Notices Amer. Math. Soc., 63(7):732–739.
- [4] I. Bárány and P. Soberón. Tverberg’s theorem is 50 years old: a survey. Bulletin of the American Mathematical Society, 55(4):459–492, 2018.
- [5] S. Bereg and M. Haghpanah. Algorithms for Radon partitions with tolerance. In 6th Internat. Conf. on Algorithms and Discrete Applied Mathematics, pages 476–487, 2020.
- [6] J. De Loera, X. Goaoc, F. Meunier, and N. Mustafa. The discrete yet ubiquitous theorems of carathéodory, helly, sperner, tucker, and tverberg. Bulletin of the American Mathematical Society, 56(3):415–511, 2019.
- [7] D. Forge, M. Las Vergnas, and P. Schuchert. 10 points in dimension 4 not projectively equivalent to the vertices of a convex polytope. European Journal of Combinatorics, 22(5):705–708, 2001.
- [8] N. García-Colín. Applying Tverberg type theorems to geometric problems. PhD thesis, University College of London, 2007.
- [9] N. García-Colín and D. G. Larman. Projective equivalences of -neighbourly polytopes. Graphs and Combinatorics, 31(5):1403–1422, 2015.
- [10] N. García-Colín, M. Raggi, and E. Roldán-Pensado. A note on the tolerant Tverberg theorem. Discrete & Computational Geometry, 58(3):746–754, 2017.
- [11] J. E. Goodman and R. Pollack. Multidimensional sorting. SIAM J. Comput., 12(3):484–507, Aug. 1983.
- [12] D. G. Larman. On sets projectively equivalent to the vertices of a convex polytope. Bulletin of the London Mathematical Society, 4(1):6–12, mar 1972.
- [13] J. Matoušek. Lectures on discrete geometry, volume 212. Springer, 2002.
- [14] W. Mulzer and Y. Stein. Algorithms for tolerant Tverberg partitions. Int. J. Comput. Geometry Appl., 24(4):261–274, 2014.
- [15] P. Soberón. Equal coefficients and tolerance in coloured Tverberg partitions. Combinatorica, 35(2):235–252, 2015.
- [16] P. Soberón. Robust Tverberg and colourful Carathéodory results via random choice. Combinatorics, Probability and Computing, 27(3):427–440, 2018.
- [17] P. Soberón and R. Strausz. A generalisation of Tverberg’s theorem. Discrete & Computational Geometry, 47(3):455–460, 2012.
- [18] H. Tverberg. A generalization of Radon’s theorem. J. London Math. Soc., 41:123–128, 1966.