通过类搜索算法实现组合测试数据集的全局优化

测试数据集的生成是软件组合测试的一个关键问题.为了提高测试数据的生成质量，提出了一种通过类搜索过程驱动的全局优化机制.在这个方法中，一个二进制编码机制被用于将组合测试数据生成问题转换为一个二进制基因序列的优化问题.同时，为了有效求解此问题，设计了一种新颖的全局优化算法—类搜索算法.此文主要论述了优化问题转换机制的可行性和有效性，并介绍了类搜索算法的计算机制.通过大量的仿真实验显示所提出的方法是可行的，且针对小规模组合测试问题，它是一种更为高效的组合测试数据集生成方法.

Global Optimization for Combination Test Suite by Cluster Searching Algorithm

The test suite generation is a key task for combinatorial testing of software. In order to generate high-quality testing data, a cluster searching driven global optimization mechanism is proposed. In this approach, a binary encoding mechanism is used to transform the combination test suite generating problem into a gene sequence optimization problem. Meanwhile, a novel global optimization algorithm, cluster searching algorithm (CSA), is developed to solve it. In this paper, the validity and rationality of problem transformation mechanism is verified, and the details of CSA are shown. The simulations illustrate the proposed mechanism is feasible. Moreover, it is a simpler and more efficient test suite generation approach for small-size combinatorial testing problems.