|
|
A351104
|
|
Numbers that begin a record-length run of consecutive numbers having the same Collatz trajectory length.
|
|
1
|
|
|
1, 12, 28, 98, 386, 943, 1494, 1680, 2987, 7083, 57346, 252548, 331778, 524289, 596310, 2886352, 3247146, 3264428, 4585418, 5158596, 5772712, 13019668, 18341744, 24455681, 98041684, 136696632, 271114753, 361486064, 406672385, 481981441, 711611184, 722067240
(list;
graph;
refs;
listen;
history;
text;
internal format)
|
|
|
OFFSET
|
1,2
|
|
COMMENTS
|
It appears that this sequence is infinite.
For every record number of consecutive identical terms in A006577, the index of the first of those consecutive terms is a term of this sequence.
This sequence is interesting because when A006577 is graphed on a scatter plot, it is immediately obvious that there are many runs of terms having the same value.
|
|
LINKS
|
|
|
EXAMPLE
|
a(4)=98 since the length of the Collatz trajectory of each number from 98 through 102 is of length 25 and this is the fourth record length.
trajectory numbers in run run
n length (1st is a(n)) length
-- ---------- -------------- ------
1 1 1 1
2 9 12, 13 2
3 18 28, 29, 30 3
4 25 98 ... 102 5
5 120 386 ... 391 6
6 36 943 ... 949 7
7 47 1494 ... 1501 8
8 42 1680 ... 1688 9
9 48 2987 ... 3000 14
10 57 7083 ... 7099 17
(End)
|
|
PROG
|
(Python)
import numpy as np
def find_records(m):
l=np.array([0]+[-1 for i in range(m-1)])
for n in range(len(l)):
path=[n+1]
while path[-1]>m or l[path[-1]-1]==-1:
if path[-1]%2==0:
path.append(path[-1]//2)
else:
path.append(path[-1]*3+1)
path.reverse()
for i in range(1, len(path)):
if path[i]<=m:
l[path[i]-1]=l[path[0]-1]+i
ciclr=[]
c=1
lsteps=0
record=0
for n in range(1, len(l)):
if l[n]==lsteps:
c+=1
else:
if c>record:
record=c
ciclr.append(n-c+1)
c=1
lsteps=l[n]
return ciclr
print(", ".join([str(i) for i in find_records(1000000)]))
|
|
CROSSREFS
|
|
|
KEYWORD
|
nonn
|
|
AUTHOR
|
|
|
STATUS
|
approved
|
|
|
|