|
|
A262725
|
|
The unique function f with f(1)=1 and f(jD!+k)=(-1)^j f(k) for all D, j=1..D, and k=1..D!.
|
|
0
|
|
|
1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1
(list;
graph;
refs;
listen;
history;
text;
internal format)
|
|
|
OFFSET
|
1
|
|
COMMENTS
|
sup_n |Sum_{j=1..n} f(jd)| is finite (but not bounded) for all d, thus giving a counterexample to a strong form of the Erdős discrepancy conjecture [see Remark 1.13 on p. 7 of Tao paper].
|
|
LINKS
|
|
|
FORMULA
|
|
|
MATHEMATICA
|
(* Generates terms recursively, using definition *)
Module[{dmax = 4, i = 2, a}, a = Table[1, (dmax+1)!]; For[d = 1, d <= dmax, d++, For[j = 1, j <= d, j++, For[k = 1, k <= d!, k++, a[[i++]] = (-1)^j*a[[k]]]]]; a] (* Paolo Xausa, Aug 10 2024 *)
(* Generates terms individually, via A034968 (slower) *)
A034968[n_] := Module[{a = n, i = 2}, While[i! <= n, a-=(i-1)*Floor[n/i++!]]; a];
|
|
PROG
|
(Sage)
A=[1, 1]
for D in [1..4]:
j=1
while j<=D:
k=1
while k<=factorial(D):
A.append((-1)^j*A[k])
k+=1
j+=1
(PARI)
A034968(n) = { my(s=0, b=2, d); while(n, d = (n%b); s += d; n = (n-d)/b; b++); (s); };
|
|
CROSSREFS
|
Cf. also A343785 (another example from Tao paper).
|
|
KEYWORD
|
easy,sign,changed
|
|
AUTHOR
|
|
|
STATUS
|
approved
|
|
|
|