A method for minimizing test sequences based on spatial-related optimization with iteration steps by set of variables

In this paper, we study structures with source-child relations projection. First we provide (deduce) a specification of several indicators and relations of subject matter. Then, we go through practical implementation. As an example, we consider a modification of algorithm proposed earlier in our research. According to our terminology, we process series of triple vectors containing variables, shorten the data by constructing a decomposition tree applying the operations of intersection, merging, and absorption of vectors within groups according to the structure. Then, we form maximally compatible subsets by a decomposition tree and perform hipping for non-specified variables by merging several vectors into one. The task is specified as a matrix of triple variables, and rows are vectors. Each vector is used in operations as a whole in spite of the vectors are series of variables. Maximally compatible subsets are considered in the scope of the given initial task by the Pareto-effective relation. Earlier, we have introduced several modifications of an algorithm, making several specifications to avoid duplication of indifferent vectors in child nodes and reducing the necessity of covering task. As usual, we determine maximally defined variable to build a next node for a decomposition tree. So, we apply the distribution by several maximally defined variables in case of a group with the same amount of defined values. From here, less steps and smaller tree are obtained. Note, an iteration approach has been proposed for the procedure, which allows avoiding excess steps and combinatory actions on the assembling stage. At this, recursion and universalization of method are used here. Also, the principle of the Layout-Related optimization is introduced in our paper.

Keywords

Authors

References