arXiv++ Combinatorics

There are several well-known methods that one can use to construct Hadamard matrices from base sequences BS(m,n). In view of the recent classification of base sequences BS(n+1,n) for n <= 30, it may be of interest to show on an example how prolific these methods are. For that purpose we have selected the Hadamard matrices of order 60. By using these methods and the transposition map we have constructed 1759 nonequivalent Hadamard matrices of order 60.

Base sequences BS(n+1,n) are quadruples of {1,-1}-sequences (A;B;C;D), with A and B of length n+1 and C and D of length n, such that the sum of their nonperiodic autocorrelation functions is a delta-function. The base sequence conjecture, asserting that BS(n+1,n) exist for all n, is stronger than the famous Hadamard matrix conjecture. We introduce a new definition of equivalence for base sequences BS(n+1,n) and construct a canonical form. By using this canonical form, we have enumerated the equivalence classes of BS(n+1,n) for n <= 30. Due to excessive size of the equivalence classes, the tables in the paper cover only the cases n <= 13.

We show that 138 odd values of n less than 10000 for which one knows how to construct a Hadamard matrix of order 4n have been overlooked in the recent handbook of combinatorial designs. There are four additional odd n, namely 191, 5767, 7081 and 8249, in that range for which we can construct a Hadamard matrix of order 4n. Our exhaustive computer searches show that the near-normal sequences NN(n) exist for n=36,38,40. Thus the Yang conjecture on the existence of NN(n) for all even n has been verified for n <= 40 but it still remains open.

The normal sequences NS(n) and near-normal sequences NN(n) play an important role in the construction of orthogonal designs and Hadamard matrices. They can be identified with certain base sequences (A;B;C;D), where A and B have length n+1 and C and D length n. C.H. Yang conjectured that near-normal sequences exist for all even n. While this has been confirmed for n not exceeding 30, so far nothing else was known for larger n. We show that NN(32) consists of 8 equivalence classes and we exhibit their representatives. We also construct representatives for two equivalence classes of NN(34). On the other hand our exhaustive computer searches show that NS(31) and NS(33) are void.

First we give an overview of the known supplementary difference sets (SDS) (A_i), i=1..4, with parameters (n;k_i;d), where k_i=|A_i| and each A_i is either symmetric or skew and k_1 + ... + k_4 = n + d. Five new Williamson matrices over the elementary abelian groups of order 25, 27 and 49 are constructed. New examples of skew Hadamard matrices of order 4n for n=47,61,127 are presented. The last of these is obtained from a (127,57,76)-difference family that we have constructed. An old non-published example of G-matrices of order 37 is also included.

We introduce a canonical form for near-normal sequences NN(n), and using it we enumerate the equivalence classes of such sequences for even n up to 30. These sequences are needed for Yang multiplication in the construction of longer T-sequences from base sequences.

The main question we raise here is the following one: given a real orthogonal n by n matrix X, is it true that there exists a rational orthogonal matrix Y having the same zero-pattern? We conjecture that this is the case and prove it for n<=5. We also consider the related problem for symmetric orthogonal matrices.

Let PCS_p^N denote a set of p binary sequences of length N such that the sum of their periodic auto-correlation functions is a delta-function. In the 1990, Boemer and Antweiler addressed the problem of constructing such sequences. They presented a table covering the range p <= 12, N <= 50 and showing in which cases it was known at that time whether such sequences exist, do not exist, or the question of existence is undecided. The number of undecided cases was rather large. Subsequently the number of undecided cases was reduced to 26 by the author. In the present note, several cyclic difference families are constructed and used to obtain new sets of periodic binary sequences. Thereby the original problem of Boemer and Antweiler is completely solved.

We construct many new cyclic (v;r,s;lambda) difference families with v less than or equal 50. In particular we construct the difference families with parameters (45;18,10;9), (45;22,22;21), (47;21,12;12), (47;19,15;12), (47;22,14;14), (48;20,10;10), (48;24,4;12), (50;25,20;20) for which the existence question was an open problem. The (45;22,22;21) difference family gives a BIBD with parameters v=45, b=90, r=44, k=22 and lambda=21, and the one with parameters (50;25,20;20) gives a pair of binary sequences of length 50 with zero periodic autocorrelation function (the periodic analog of a Golay pair). We also construct nine new D-optimal designs. A normal form for cyclic difference families is proposed and used effectively in compiling the list of known and new difference families.

We construct two difference families on each of the cyclic groups of order 109, 145 and 247, and use them to construct skew-Hadamard matrices of orders 436, 580 and 988. Such difference families and matrices are constructed here for the first time. The matrices are constructed by using the Goethals-Seidel array.

We construct several difference families on cyclic groups of orders 47 and 97, and use them to construct skew-Hadamard matrices of orders 188 and 388. Such difference families and matrices are constructed here for the first time. The matrices are constructed by using the Goethals-Seidel array.

We construct two Hadamard matrices of order 764. Both are of Goethals-Seidel type.