A Method of Testing Generation Based on Prioritized Pair

Combinatorial testing has lots of test cases, but software testers hope to get the best test coverage with the smallest test case suite. For the scale of produced test cases is so large that researchers have considered the implementation of the critical test cases. This article researches the classic combinatorial test methods and proposes methods to generate pair-wise testing cases with a priority. Firstly, we design formulas to compute the weights of priorities. Secondly, we adopt a greed algorithm to solve the combinatorial testing problems. Furthermore, we integrate the greed strategy into a genetic algorithm to improve the efficiency. It improves the testing efficiency while securing the detection rate of defects under limited resources.

Real-Space Descriptions of Structural Energies in Metals

ABSTRACTTwo real-space methods for treating structural energetics of transition metals and compounds are described. The first uses a local description of the electronic density of states (DOS) in a tight-binding model to obtain an angular-force method containing up to four-body interaction terms. It is shown that this method yields bond-strengthening effects at surfaces which exceed those obtained by previous many-body potentials. Structural-energy calculations show that for W, several Frank-Kasper phases are only slightly higher in energy than the ground-state bcc structure; this near-degeneracy is driven by the energetic favorability of icosahedral sites. The second method uses a free-electron type approach to generate pair potentials for transition-metal solutes in Al. The calculated potentials have an oscillating form, with a much larger magnitude than those in Al. The potential is applied to complex Al-Mn phases, including the icosahedral quasicrystal. The results indicate that the oscillating pair potentials make a major contribution to stabilizing the complex phases.