The maximum distinct differences problem asks for the largest number of distinct positive differences
determined by a
-element subset of the integers
0, 1, ...,
,
where
is a nonnegative integer and
. Explicitly,
|
(1)
|
where the maximum is over the stated -element subsets. Pikhurko and Schoen
(2007, pp. 1-2) use
for the largest cardinality
of
,
including zero and both signs, so
. The number of unordered pairs and the available
differences give
|
(2)
|
Attaining the first bound requires all pairwise differences to be distinct, as in a Golomb ruler. Attaining the second requires every
distance from 1 through to occur. A sparse ruler minimizes
the number of marks needed for that complete coverage, whereas this problem fixes
the number of marks and maximizes the number of distances attained.
A sequence of particular interest takes , where the two elementary bounds agree,
|
(3)
|
Beginning with , the recorded values are 3, 6, 9, 13, 19, 26, 32, 40, 49,
58, 68, 79, ... (OEIS A399997). This sequence
was a runner-up in the OEIS 400K celebration contest (OEIS Foundation Inc. 2026).
For ,
the marks
0,
1, 3, 13, 20, 27, 31, 35, 36
attain 32 distances, missing only 6, 21, 25, and 29 (Pegg 2026).
Adding a mark at 6 gives all 36 distances, illustrating the distinction between maximizing
partial coverage and requiring complete coverage. Pegg (2026) also supplies larger
constructions. For example, his 15-mark construction attains 90 distinct positive
differences and therefore establishes
, with equality conjectural.
For any fixed , Pikhurko and Schoen (2007, pp. 3-4) proved that
|
(4)
|
exists whenever , independently of the choice of the integer-valued function
.
Here
denotes a quantity tending to zero. Consequently, the special sequence
satisfies
|
(5)
|
This limit concerns asymptotic density and does not establish exact finite values.