OFFSET
1,1
COMMENTS
Inspired by a comment of Mario Cortés in A090897, who suggests that 1 might not appear in this sequence.
Differs from A336519 for n = 4, 16, 73, 83, 90, ....
a(n) is not 1 for the first 2000 terms. We can prove that a(n) has a prime factor p that does not divide LCM(A090897(1), ..., A090897(n-1)) without using prime number factorization. The method is explained in the link below. - David A. Corneth, Aug 22 2020
LINKS
EXAMPLE
[ 1] 3, {3} -> 3;
[ 2] 14, {2, 7} -> 2;
[ 3] 159, {3, 53} -> 53;
[ 4] 2653, {7, 379} -> 379;
[ 5] 58979, {58979} -> 58979;
[ 6] 323846, {2, 161923} -> 161923;
[ 7] 2643383, {2643383} -> 2643383;
[ 8] 27950288, {2, 1746893} -> 1746893;
[ 9] 419716939, {6971, 60209} -> 6971;
[10] 9375105820, {2, 5, 1163, 403057} -> 5.
PROG
(SageMath)
def Select(item, Selected):
return next((x for x in item if not (x in Selected)), 1)
def PiPart(n):
return floor(pi * 10^(n * (n + 1) // 2 - 1)) % 10^n
def A336520List(len):
prev = []; ret = []
for n in range(1, len + 1):
p = prime_factors(PiPart(n))
ret.append(Select(p, prev))
prev.extend(p)
return ret
print(A336520List(39))
# Query function of David A. Corneth to determine if a(n) is prime.
def LcmPiPart(n):
return lcm([PiPart(n) for n in (1..n)])
def is_an_prime(n):
lcmpi = LcmPiPart(n - 1)
lm, m = 1, PiPart(n)
while lm != m:
lm, m = m, lcm(lcmpi, m) // lcmpi
return m > 1
CROSSREFS
KEYWORD
nonn,base
AUTHOR
Peter Luschny, Aug 22 2020
STATUS
approved