On t-sumfree sequences
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: We're looking at the paper: "On t-sumfree sequences".
Emmy: Here's the gist of the paper. In this paper we explore the world of t-sumfree sequences, primarily for t = 2 and t = 3.
Ada: The main open problem concerns the question of the conjectured ultimate periodicity of these sequences.
Emmy: Definition and Context The infinite sequences of increasing positive integers we study in this paper arise as follows.
Ada: The requirement that sn is the smallest integer is the greedy aspect, and that it is not the sum of t different previous entries is the strict part.
Paper summary: Ada: So, looking at "On t-sumfree sequences," the authors have thoroughly explored the world of these greedy sequences for t being two and three.
Emmy: The main point they push is that while proving ultimate periodicity for all cases is tough, their results show that a large chunk of these sequences are indeed ultimately periodic, with only a small, finite number of exceptions identified under specific parameter constraints.
Kolya: It settles the conjecture about eventual periodicity for three-sumfree sequences to be true, based on the extensive computational experiments they ran and the theorems they derived.
Ada: Essentially, it means that if you're building these sequences greedily, you don't have to worry about them going on forever without repeating; there's a predictable cycle waiting for you after some initial setup.
Emmy: The title itself suggests a deep dive into the nature of these sequences, and the results from this paper give us strong evidence for that deep structure existing in their behavior.
Kolya: It’s about confirming that even though they are constructed by simple rules, there is an underlying periodicity governed by those constraints, and that's what the paper is really pointing toward.
Conclusion: Ada: So we've been looking at these sequences called t-sumfree sequences for t being two and three, specifically focusing on how they behave as they get longer and what that tells us about their ultimate structure.
Emmy: The paper we just read, "On t-sumfree sequences," it really zeroes in on those greedy rules—where the next number has to be the smallest possible integer but can't be made by adding up any of the previous numbers in a certain way.
Ada: And what they’re really digging into is this big question about ultimate periodicity. They have these specific results, like proving periodicity for a big chunk of two-sumfree and three-sumfree sequences, but the main thrust is figuring out if *all* of them settle into a repeating pattern eventually.
Emmy: It seems they’ve managed to prove that this conjecture is true by showing that even when things look complicated, like in the case for three-sumfree sequences, there are only a finite number of exceptions to the rule.
Ada: That’s what Theorem five is all about—limiting those exceptions to just twelve possibilities under certain constraints on the initial numbers. It feels like they've turned a very open conjecture into something provable for most cases.
Emmy: The authors use extensive computation, looking at thousands of examples, and then use that data to back up their theorems, which is pretty neat because it bridges the gap between what we think should be true and what we can actually check.
Ada: It makes you wonder how this applies outside of pure number theory. If these sequences are predictable in the long run, maybe that structure shows up in other areas where patterns matter, like cryptography or even some aspects of data compression.
Emmy: Exactly. The core idea is that these simple greedy rules generate deep structure, and by finding those periodic bounds, we’re mapping out the long-term behavior of these number sequences.
Ada: But there are still some specific families left over, and figuring out exactly what those exceptions look like is still a major piece of the puzzle for them.
Emmy: That’s where we head next; we need to see those specific cases they're worried about, because that’s where the real challenge lies.
math.NT
Submitted: 2026-09-15
Updated: 2026-11-11
Comments: 41 pages, 3 tables
Code: https://github.com/fifth-postulate/subsumfree
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 84/100
The gist: In this paper we explore the world of t-sumfree sequences, primarily for t = 2 and t = 3.
Key concepts
- t-sumfree sequence
- These are infinite sequences of increasing positive integers where no term is the sum of t different preceding terms. The requirement ensures a strict constraint on how terms can be generated from previous elements in the sequence.
- Ultimate Periodicity
- This refers to whether an infinite sequence eventually settles into a repeating pattern after an initial, finite segment (the preperiod). The main conjecture explored is that all t-sumfree sequences are eventually periodic.
- Greedy Aspect
- This aspect of sequence generation means that each term in the sequence must be the smallest possible positive integer that satisfies the sumfree condition based on previous terms. It dictates a specific, deterministic way to build the infinite sequence.
- Preperiod and Period
- When a sequence is ultimately periodic, it consists of an initial non-repeating segment called the preperiod, followed by an infinitely repeating block called the period. The paper seeks to find conditions that define these exact lengths for specific sequences.
Terminology
Summary
In this paper we explore the world of t-sumfree sequences, primarily for t = 2 and t = 3. p. 0 ON t-SUMFREE SEQUENCES
DAAN VAN BERKEL AND WIEB BOSMA
Abstract. In this paper we explore the world of t-sumfree sequences, primarily for t = 2 and t = 3. Starting with t increasing positive integers, build
an infinite sequence by choosing as the next entry the least positive integer
that is not the sum of t distinct previous values. The main open problem concerns the question of the conjectured ultimate periodicity of these sequences.Where the paper says it
Where the paper says it
Definition and Context
The infinite sequences of increasing positive integers we study in this paper arise as follows. The requirement that sn is the smallest integer is the greedy aspect, and that it is not the sum of t different previous entries is the strict part. p. 0 Introduction
The infinite sequences of increasing positive integers we study in this paper
arise as follows: initialize the first t ≥ 1 values s1 < s2 < · · · < st, and for
n > t let sn be the smallest integer exceeding sn−1 that is not equal to the sum
of t different previous elements in the sequence; these are greedy, strict t-sumfree
sequences, which we will refer to in what follows in short as t-sumfree sequences. The requirement that sn is the smallest integer is the greedy aspect, and that it is
not the sum of t different previous entries is the strict part. The requirement that sn is the smallest integer is the greedy aspect, and that it is
not the sum of t different previous entries is the strict part. The main new results
in this paper, which we will summarize in this introduction, concern the case t = 3,
and the most interesting problem is the periodicity question.Where the paper says it
Where the paper says it
Periodicity Conjectures and Results
The main conjecture is that all t-sumfree sequences are ‘eventually periodic’. The authors formulate a sufficient condition to prove that a conjectured preperiod and period are correct. p. 0 The main open problem concerns the question of the conjectured ultimate periodicity of these sequences. We formulate a sufficient condition to prove that a conjectured preperiod and
period are correct. We formulate a sufficient condition to prove that a conjectured preperiod and
period are correct. We also prove periodicity for a considerable subclass of 2-
sumfree and of 3-sumfree sequences. These experiments lead to a proof of the
following theorem. Among all triples (f, g, h) with f < g < h, satisfying f ≤ 25 and
d = h − g ≤ 25 there are at most 12 exceptions to the rule that the 3-sumfree
sequence Sf,g,h is ultimately periodic.Where the paper says it
Where the paper says it
Where the paper says it
Characterizations of Specific Sequences
The paper presents several theorems characterizing specific families of sequences. For instance, Theorem 13 characterizes the 2-sumfree sequence Sf,f+d as follows. p. 8 Theorem 16. For every g ≥ 2 the 3-sumfree sequence S1,g,g+1 is characterized as
follows:
z ∈ S1,g,g+1 ⇐⇒ z ∈ {1, 2g + 1, 6g + 1} or
z ≡ {g, g + 1, · · · , 2g} ∪ {6g + 2, 6g + 3, · · · , 7g + 1} mod 10g + 3. For g ≥ 8 the greedy 2-sumfree sequence S1,g is
characterized as follows:
z ∈ S1,g ⇐⇒ z ∈ {1, g, g + 2, g + 4, · · · , 2g, 2g + 3} or z > 2g + 4 and
z ≡ {2g + 5, 2g + 7, 2g + 9, · · · , 3g + 1, 4g, 4g + 2, 4g + 4,
4g + 6, 5g + 3, 5g + 5, 5g + 7, · · · , 6g + 1} mod 4g + 3. In particular, for g ≥ 8 after the first e + 3 entries (in a preperiod) the sequence
S1,g modulo 4g + 3 is periodic with period length g + 3. For any odd g the greedy 2-sumfree sequence S1,g is given by
z ∈ S1,g ⇐⇒ z = 1 or z ≥ g and odd. Thus, S1,2e+1 is ultimately periodic modulo 2, with preperiod length 1 and period
length 1, except for S1,3 which is periodic from the beginning.Where the paper says it
Where the paper says it
Where the paper says it
Computational Verification
Fairly extensive computations do lead us to believe that Conjecture 4, that every 3-sumfree sequence Sf,g,h is ultimately periodic, is true. The authors use Lemma 6 to show how such proofs follow from a finite computation. p. 2 All but 12 of
these were shown to be ultimately periodic. For the proofs of ultimate periodicity,
we used an elementary lemma (Lemma 6 in Section 3) that shows how such proofs
follow from a finite computation. ON t-SUMFREE SEQUENCES 3
Note that for every pair of parameters (f, d), Theorem 3 leaves only finitely many
3-sumfree sequences Sf,g,g+d (values of g) that could possibly be not ultimately
periodic. We did very extensive experiments to provide evidence for the next conjecture. Among all triples (f, g, h) with f < g < h, satisfying f ≤ 25 and
d = h − g ≤ 25 there are at most 12 exceptions to the rule that the 3-sumfree
sequence Sf,g,h is ultimately periodic. Theorem 3 leaves at most 2024 triples (f, g, h) with f ≤ 25 and d ≤ 25 to be
investigated: for each d ≥ 4 there are (d−3)(d−2)
2
triples not covered.Where the paper says it
Where the paper says it
Where the paper says it
Algorithmic Tools
The paper discusses algorithms for computing sumfree sequences. It suggests using a heap data structure to retrieve the combination with the smallest sum when computing the next term. p. 4 With the aid of a heap data structure ([9], 5. 3),
it is not difficult to retrieve the combination with the smallest sum, using that S
is strictly increasing. All but 12 of
these were shown to be ultimately periodic. For the proofs of ultimate periodicity,
we used an elementary lemma (Lemma 6 in Section 3) that shows how such proofs
follow from a finite computation. Determining periods. To prevent computing unnecessary large initial segments, it is good to use Lemma 6 to check if period and preperiod can be obtained
from a given initial segment. The main open problem concerns the question of the conjectured ultimate periodicity of these sequences. We formulate a sufficient condition to prove that a conjectured preperiod and
period are correct. The reservations we have about it, stem from our attempts to provide computational evidence for this conjecture. These experiments lead to a proof of the
following theorem. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Related Work
There is a considerable literature on sumfree sets and sequences. Dickson was the first to coin questions about sumfree sequences. p. 2 Apparently (cf. Guy, [8] Problem E32) Dickson [7] was the first to coin questions
about sumfree sequences; however, his main focus was on constructing sequences
with the property that every positive integer is the sum of (at most) t elements in
the sequence (allowing repetitions). Guy, [8] Problem E32) Dickson [7] was the first to coin questions
about sumfree sequences; however, his main focus was on constructing sequences
with the property that every positive integer is the sum of (at most) t elements in
the sequence (allowing repetitions). Queneau [11] defined s-additive sequences, which consist, after an initial segment,
of those positive integers that can be expressed in exactly s ways as the sum of two
different previous entries. 4 VAN BERKEL AND BOSMA
Cameron noted a natural bijection between the set of non-strict, greedy 2-sumfree
sequences and binary sequences σ. For given σ the positive integers n are tested
in order for inclusion in S: when n is not the sum of two elements of S the first
element of σ not used yet decides whether n is added to S (if it is 1) or not. Cameron [6] proved that S is ultimately periodic if σ is, but the converse is not
clear. In fact Calkin and Finch [4] did extensive calculations which left several
periodic candidates for σ that seem to give aperiodic S. In the case of sumfree sets there is often emphasis on questions of density: how
large can sumfree sets be relative to their universe? For our sequences we only
mention that it seems reasonable to conjecture that the maximal density for any
t-sumfree sequence is 1
t
, compare Theorem 9 and Theorem 13. The reservations we have about it, stem from our attempts to provide computational evidence for this conjecture. These experiments lead to a proof of the
following theorem. The reservations we have about it, stem from our attempts to provide computational evidence for this conjecture. These experiments lead to a proof of the
following theorem.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Computational Results Summary
Table 1 lists small cases, and Table 2 summarizes all exceptionally large values for period and preperiod of Sf,g,h that occurred for all 2024 triples (f, g, h) with 1 ≤ f < g < h for which both f ≤ 2 and d = h − g ≤ 25. An entry with k = p = −1 signifies that no periodicity was found at 500000 terms. p. 2 For the proofs of ultimate periodicity,
we used an elementary lemma (Lemma 6 in Section 3) that shows how such proofs
follow from a finite computation. Section 4 contains some remarks on the algorithms we implemented, and Section
7 contains some computational details, such as tables with (parts of) the results
mentioned in Theorem 5 and the number of terms (at least 500000) computed for
the 12 possible exceptions. In the above example this characteristic
sequence C = C1,2,3 starts as
C = (1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, · · ·),
where the final 1 here refers to the entry 27 in S1,2,3. A consequence of such
ultimate periodicity is that there exists a positive integer m such that S modulo m
is periodic from a certain entry on, with period length that we will indicate by p
Date: today.Where the paper says it
Where the paper says it
Exceptions and Limitations
Theorem 19 leaves only finitely many 3-sumfree sequences Sf,g,g+d (values of g) that could possibly be not ultimately periodic. The reservations about Conjecture 4 stem from computational evidence. p. 2 Every 3-sumfree sequence Sf,g,h is ultimately periodic. The reservations we have about it, stem from our attempts to provide computational evidence for this conjecture. Then for every f ≥ 1
and every g ≥ f + 1 such that f + g ≥ d + 1 the greedy 3-sumfree sequence Sf,g,g+d
is ultimately periodic modulo 5g + 2d+ 3(f −1), with preperiod length g +f + 2 and
period length g + f. More explicitly, Df,g,g+d is ultimately periodic in these cases,
with preperiod
(g − f, d, 1
f+g−1
, 4g + 2f + 2(d − 2))
and period
(1f+g−1
, 4g + 2f + 2(d − 1)). □
The situation for even g is slightly more complicated, and is most easily described
using the sequence of differences. Theorem 10. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly. 8 VAN BERKEL AND BOSMA
with 8 and the residue classes of 1, 4, 6, 11 modulo 12; and
S1,6 = 1, 6, 8, 10, 12, 15, 17, 19, 24, 26, 28, 33, 35, 37, 42, 44, 46, 51, 53, · · ·
containing 12 and the complete residue classes 1, 6, 8 mod 9. The next theorem describes another infinite family, which we mainly give for
comparison with the case t = 3: it turns out that families of this kind, with fixed
difference between two parameters, are easier to describe explicitly.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Implement a verification algorithm based on Lemma 6 to formally prove ultimate periodicity for any given t-sumfree sequence by checking if a non-zero block of length M repeats sufficiently often according to the condition (Q - t(t+1)/2)M >= (t - 1)K.
-
Develop an algorithmic routine that uses the characteristic sequence C or the first difference sequence Df,g,h to efficiently determine the preperiod length k and period length p by applying Corollary 7 or Lemma 6 checks based on a finite initial segment computation.
-
Create a systematic computational exploration framework that leverages Theorem 5 to reduce the search space for non-periodic triples (f, g, h) by only investigating cases where d <= 25 and f <= 25, thereby ensuring all known exceptional cases are rigorously tested.
-
Implement a function that characterizes specific families of sequences using Theorem 13 for t=2 and t=3, allowing the system to predict the modulus m and period length p based on input parameters f and d.
-
Integrate a mechanism to handle large initial segments by utilizing caching of partial sums when computing next terms, as suggested in Section 4.1, to avoid redundant calculations during sequence generation up to a bound.
Abstract
In this paper we explore the world of t-sumfree sequences, primarily for t=2 and t=3. Starting with t increasing positive integers, build an infinite sequence by choosing as the next entry the least positive integer that is not the sum of t distinct previous values. The main open problem concerns the question of the conjectured ultimate periodicity of these sequences. We formulate a sufficient condition to prove that a conjectured preperiod and period are correct. We also prove periodicity for a considerable subclass of 2-sumfree and of 3-sumfree sequences. Finally, we present the results of extensive computations on 3-sumfree sequences.
Related papers
- K ab has the Bogomolov property for canonical heights
- Quasi-Jacobi forms, Appell-Lerch functions, and false theta functions as q-brackets of functions on partitions
- Wasserstein rates of all orders for the Sato-Tate law of Kloosterman sums
- Completing the classification of torsion over sextic fields for elliptic curves defined over Q
- Riemann functions for almost primes and their counting errors
- Topological flatness of orthogonal spin local models