一类可分解的Mendelsohn设计的存在性.pdfVIP

Introduction In the history of attempts to perform scientifically sound experiments, com- binatorial design, which study the combinatorial question that arose originally from issues in experimental design, has played an important role. We can apply combinatorial design to topics in cryptography, design of computer and com- munication networks, software testing, storage in disk arrays, signal process- ing, sports scheduling, designing chips for DNA probes, and clone screening in molecular biology. The existence problem, which deals with the question: there is whether at least one arrangements of a particular kind or not, is a basic problem of combi- natorial design. We shall illustrate the problem with a example of experimental design. Let us now introduce some general terminology. Suppose that P is a set of experimental units or plots, and V is a set of treatments or varieties. Certain subsets of P will be called blocks. Given P and V, a block design is defined by giving the collection of blocks and assigning to each experimental unit in P a treatment in V. Thus corresponding to each block is a set of treatments. Speaking abstractly, we shall be able to disregard the experimental units and think of a block design as simply consisting of a set V of treatments and a collection of subsets of V, called blocks. To be concert, let us consider an experiment designed to test the effect on hu- man beings of five different drugs. Let the drugs be labelled 1, 2, 3, 4, 5. We could pick out five subjects and give each subject a different drug. Unfortunately, cer- tain subjects might be allergic to a particular drug or immune to its effects. Thus, we could get very biased results. A more effective use of five subjects would be to 1 give each of the drugs, say on five consecutive days. I