cp's OEIS Frontend

The terms for row n are computed as A227451(n), A226062(A227451(n)), A226062(A226062(A227451(n))), etc. until a term that is a fixed point of A226062 is reached (A037481(n)), which will be the last term of row n.

Row n has A002061(n) = 1,1,3,7,13,21,... terms.

Factor n; replace each prime(i) with i, take the conjugate partition, replace parts i with prime(i) and multiply out.

From Antti Karttunen, May 12-19 2014: (Start)

For all n >= 1, A001222(a(n)) = A061395(n), and vice versa, A061395(a(n)) = A001222(n).

Because the partition conjugation doesn't change the partition's total sum, this permutation preserves A056239, i.e., A056239(a(n)) = A056239(n) for all n.

(Similarly, for all n, A001221(a(n)) = A001221(n), because the number of steps in the Ferrers/Young-diagram stays invariant under the conjugation. - Note added Apr 29 2022).

Because this permutation commutes with A241909, in other words, as a(A241909(n)) = A241909(a(n)) for all n, from which follows, because both permutations are self-inverse, that a(n) = A241909(a(A241909(n))), it means that this is also induced when partitions are conjugated in the partition enumeration system A241918. (Not only in A112798.)

(End)

From Antti Karttunen, Jul 31 2014: (Start)

Rows in arrays A243060 and A243070 converge towards this sequence, and also, assuming no surprises at the rate of that convergence, this sequence occurs also as the central diagonal of both.

Each even number is mapped to a unique term of A102750 and vice versa.

Conversely, each odd number (larger than 1) is mapped to a unique term of A070003, and vice versa. The permutation pair A243287-A243288 has the same property. This is also used to induce the permutations A244981-A244984.

Taking the odd bisection and dividing out the largest prime factor results in the permutation A243505.

Shares with A245613 the property that each term of A028260 is mapped to a unique term of A244990 and each term of A026424 is mapped to a unique term of A244991.

Conversely, with A245614 (the inverse of above), shares the property that each term of A244990 is mapped to a unique term of A028260 and each term of A244991 is mapped to a unique term of A026424.

(End)

The Maple program follows the steps described in the first comment. The subprogram C yields the conjugate partition of a given partition. - Emeric Deutsch, May 09 2015

The Heinz number of the partition that is conjugate to the partition with Heinz number n. The Heinz number of a partition p = [p_1, p_2, ..., p_r] is defined as Product(p_j-th prime, j=1...r). Example: a(3) = 4. Indeed, the partition with Heinz number 3 is [2]; its conjugate is [1,1] having Heinz number 4. - Emeric Deutsch, May 19 2015

Like A129594 this sequence utilizes the fact that compositions (i.e., ordered partitions) can be bijectively mapped to (unordered) partitions by taking the partial sums of the list of composants after one has been subtracted from each except the first one. Compositions in turn are mapped to nonnegative integers via the runlength encoding, where the lengths of maximum runs of 0's or 1's in binary representation of n give the composants. See the OEIS Wiki page and the example below.

Each n occurs A000041(n) times in total and occurs for the first time at A227368(n) and for the last time at position A000225(n). See further comments and conjectures at A227368 and A227370.

For this sequence the partitions are encoded in the binary expansion of n in the same way as in A129594.

In "Bulgarian solitaire" a deck of cards or another finite set of objects is divided into one or more piles, and the "Bulgarian operation" is performed by taking one card from each pile, and making a new pile of them. The question originally posed was: on what condition the resulting partitions will eventually reach a fixed point, that is, a collection of piles that will be unchanged by the operation. See Martin Gardner reference and the Wikipedia-page.

A037481 gives the fixed points of this sequence, which are numbers that encode triangular partitions: 1 + 2 + 3 + ... + n.

A227752(n) tells how many times n occurs in this sequence, and A227753 gives the terms that do not occur here.

Of further interest: among each A000041(n) numbers j_i: j1, j2, ..., jk for which A227183(j_i)=n, how many cycles occur and what is the size of the largest one? (Both are 1 when n is in A000217, as then the fixed points are the only cycles.) Cf. A185700, A188160.

Also, A123975 answers how many Garden of Eden partitions there are for the deck of size n in Bulgarian Solitaire, corresponding to values that do not occur as the terms of this sequence.

The terms have a particular pattern in their binary expansion, which encodes for a "triangular partition" when runlength encoding of unordered partitions are used (please see A129594 for how that encoding works).

n a(n) same in binary run lengths unordered partition

0 0 0 [] {}

1 1 1 [1] {1}

2 6 110 [2,1] {1+2}

3 25 11001 [2,2,1] {1+2+3}

4 102 1100110 [2,2,2,1] {1+2+3+4}

5 409 110011001 [2,2,2,2,1] {1+2+3+4+5}

6 1638 11001100110 [2,2,2,2,2,1] {1+2+3+4+5+6}

7 6553 1100110011001 [2,2,2,2,2,2,1] {1+2+3+4+5+6+7}

8 26214 110011001100110 [2,2,2,2,2,2,2,1] {1+2+3+4+5+6+7+8}

9 104857 11001100110011001 [2,2,2,2,2,2,2,2,1] {1+2+3+4+5+6+7+8+9}

These partitions are the only fixed points of "Bulgarian Solitaire" operation (see Gardner reference or Wikipedia page), and thus the terms of this sequence give the fixed points for A226062 which implements that operation (using the same encoding for partitions). This also implies that these partitions are the roots of the game trees constructed for decks consisting of 1+2+3+...+k cards. See A227451 for the encoding of the corresponding tops of the main trunks of the same trees. - Antti Karttunen, Jul 12 2013

a(0) = 1, as 0 is here considered to encode an empty partition {}, and the empty product is one.

Like A129594, this sequence is based on the fact that compositions (i.e., ordered partitions) can be mapped 1-to-1 to partitions by taking the partial sums of the list where one is subtracted from each composant except the first (originally explained by Marc LeBrun in his Jan 11 2006 post on SeqFan mailing list, with an additional twist involving factorization and prime exponents, cf. A129595). The example below show how this works.

Compare the scatterplot of this sequence to those of A002487, A243353, A243499 and A253552.

As described by Marc LeBrun, we can map integers 1-to-1 to partitions in a "crazy" order: factor n, take the (finite) tuple of exponents, add 1 to the first, use the rest as successive differences between parts and finally subtract 1 from the last part, thus we get the following partitions (elements in ascending order): 2 -> [1] -> 1, 3 -> [0,1] -> 1+1, 4 -> [2] -> 2, 5 -> [0,0,1] -> 1+1+1, 6 -> [1,1] -> 2+2, 7 -> [0,0,0,1] -> 1+1+1+1, 8 -> [3] -> 3, 9 -> [0,2] -> 1+2, 10 -> [1,0,1] -> 2+2+2, etc.

Inverse process: from a sorted (elements in ascending order) partition of n, subtract 1 from the first part, then take the first differences of parts and add 1 to the last (of differences or the first part if a singular partition) and use them as the exponents for A000040(1), A000040(2), etc. and multiply.

This array is obtained when we encode in such a way the partition obtained as an element-wise sum of two partitions encoded by i and j. The element-wise addition begins from the largest elements of the partitions, continuing towards the smaller elements and if the partitions do not contain the same number of elements, the shorter is prepended with as many zeros as needed to make them of equal length.

On what condition does A(i,j) = i*j ? E.g., A(3,5)=15, A(3,10)=30, A(5,11)=55. However A(3,7)=35 and A(5,7)=21.

This is a self-inverse permutation of the positive integers.

Old name was: a(0) = 0, and for n>=1, let b(n,m) be the m-th digit, reading left to right, of binary n. (b(n, 1) is the most significant binary digit, which is 1.) Then a(n) is such that b(a(n),1)=1; and if b(n,m)=b(n,m-1) then b(a(n),m) does not = b(a(n),m-1); and if b(n,m) does not = b(n,m-1) then b(a(n), m) = b(a(n),m-1), for all m where 2 <= m <= number binary digits in n.

From Emeric Deutsch, Oct 06 2020: (Start)

a(n) is the index of the composition that is the conjugate of the composition with index n.

The index of a composition is defined to be the positive integer whose binary form has run-lengths (i.e., runs of 1's, runs of 0's, etc. from left to right) equal to the parts of the composition. Example: the composition 1,1,3,1 has index 46 since the binary form of 46 is 101110.

a(18) = 24. Indeed, since the binary form of 18 is 10010, the composition with index 18 is 1,2,1,1 (the run-lengths of 10010); the conjugate of 1,2,1,1 is 2,3 and so the binary form of a(18) is 11000; consequently, a(18) = 24. (End)

The bijective encoding of nonordered partitions via compositions (ordered partitions) present in the binary expansion of n is explained in A227184.

It appears that a(4n+2) = a(2n+1). - Ralf Stephan, Jul 20 2013

The terms have particular patterns in their binary expansion, which encodes for an "almost triangular partition" when runlength encoding of unordered partitions are used (please see A129594 for how that encoding works). These are obtained from the perfectly triangular partitions shown in A037481 by inserting 1 to the front of the partition and decrementing the last summand (the largest) by one:

n a(n) same in binary run lengths unordered partition

0 0 0 [] {}

1 1 1 [1] {1}

2 5 101 [1,1,1] {1+1+1}

3 18 10010 [1,2,1,1] {1+1+2+2}

4 77 1001101 [1,2,2,1,1] {1+1+2+3+3}

5 306 100110010 [1,2,2,2,1,1] {1+1+2+3+4+4}

6 1229 10011001101 [1,2,2,2,2,1,1] {1+1+2+3+4+5+5}

7 4914 1001100110010 [1,2,2,2,2,2,1,1] {1+1+2+3+4+5+6+6}

8 19661 100110011001101 [1,2,2,2,2,2,2,1,1] {1+1+2+3+4+5+6+7+7}

9 78642 10011001100110010 [1,2,2,2,2,2,2,2,1,1] {1+1+2+3+4+5+6+7+8+8}

These partitions occur at the tops of the main trunks of the game trees constructed for decks consisting of 1+2+3+...+k cards. See A037481 for the encoding of the roots of the main trunks of the same trees.