An extending theorem for \(s\)-resolvable \(t\)-designs (Q2657143)

Given a \(v\)-set \(X\) of points and a collection \(\mathcal{B}\) of mutually distinct \(k\)-subsets of \(X\), called blocks, such that every \(t\)-subset of \(X\) is contained in exactly \(\lambda_t\) blocks of \(\mathcal{B}\), the pair \((X, \mathcal{B})\) is called a \(t\)-\((v, k, \lambda_t )\) simple design. If \(B\) is the set \(\binom{X}{k}\) of all \(k\)-subsets of \(X\), then \((X, \mathcal{B})\) is a \(t\)-\((v, k, \binom{v-t}{k-t})\) design, called the complete design. A \(t\)-\((v, k, \lambda )\) simple design \((X, \mathcal{B})\) is said to be \(s\)-resolvable, for \(0 < s < t\), if its block set \(\mathcal{B}\) can be partitioned into \(N \geq 2\) classes \(A_1,\dots, A_N\) such that each \((X, A_i)\) is an \(s\)-\((v, j, \delta)\) design for \(i = 1,\dots, N\). Each \(A_i\) is called a resolution class and the set of \(N\) classes is called an \(s\)-resolution of \((X, \mathcal{B})\). A \(t\)-resolution \(\{(X, A_i)\}_{i=1}^N\) with \(N\) resolution classes of the complete \(k\)-\((v, k, 1)\) design \((X,\binom{X}{k})\), each class being a \(t\)-\((v, k, \lambda )\) design, is called a large set \(LS[N](t, k, v)\). The author proves that, if \(D_1\) and \(D_2\) are simple \(t\)-\((v_1, k, \lambda_1)\) and \(t\)-\((v_2, k, \lambda_2)\) designs, with \({\lambda_1}=z{\binom{v_1-t}{k-t}}\) and \({ \lambda_2}=z{\binom{v_2-t}{k-t}}\) for some positive \(z\in \mathbb{Z}\), such that \(D_1\) and \(D_2\) are both \(s\)-resolvable with \(N\) resolution classes and \(z = N\frac{u}{n}\) for positive integers \(u\) and \(n\), and if there exist \(LS[n](k - 2, k - 1, v_1 - 1)\) and \(L S[n](k - 2, k - 1, v_2 - 1)\), then there exists a simple \(s\)-resolvable \(t\)-\((v_1 +v_2 -k +1, k, \lambda)\) design \(D_3\) with \(N\) resolution classes, where \(\lambda = z\binom{v_1+v_2-k+1-t}{k-t}\). This extends significantly a theorem in [\textit{Q.-r. Wu}, Australas. J. Comb. 4, 229--235 (1991; Zbl 0763.05013)], and applies to a series of corollaries and examples, among which we remark the following: 1) There exist simple \(3\)-resolvable \(4\)-designs with \(N = 5\) resolution classes having parameters \(4\)-\((34, 6, 75)\) and \(4\)-\((35, 7, 31 \cdot 25)\). 2) There exists a \(2\)-resolvable \(3\)-\((3 + 29m, 4, 5m)\) design with \(N = 5\) resolution classes for any \(m \geq 1\). 3) For any given integer \(h \geq 1\), there exist \(N = 5\cdot 2h -5\) mutually disjoint simple \(3\)-\((3 + m(5 \cdot 2h - 3), 4, m)\) designs for all \(m \geq 1\). The author also leaves an open problem, whose positive answer would directly give in turn the existence of simple designs with interesting parameters.