论文部分内容阅读
软件产品线工程是一个利用更全面的软件复用技术能够在开发成本和产品上市时间方面极大改善软件开发过程的有效方式.特征模型是软件产品线工程中描述软件产品家族共性和可变性特征以及特征之间关系的需求模型.基于特征模型的产品配置难点在于如何从复杂的特征模型中选择满足约束条件的最优的特征组合,如资源约束等.为了实现有约束的特征选择优化问题,提出一种基于原子集和遗传算法优化特征选择的方法.该方法是在原子集优化特征模型基础上,利用特征模型的整体约束信息将种群分为有效配置和无效配置,对有效配置和无效配置进行交叉等遗传操作使其加速向最优解收敛.实验结果说明了该方法的有效性.
Software Product Line Engineering is an effective way to dramatically improve the software development process with more comprehensive software reuse techniques that characterize the family commonality and variability of software products in software product line engineering As well as the requirements of the relationship between the characteristics of the model.The product model based on the characteristics of the difficulty of configuration is how to select from the complex feature model to meet the constraints of the optimal combination of features such as resource constraints.In order to achieve constrained feature selection optimization problems, This paper proposes a method of optimizing feature selection based on atomic subset and genetic algorithm. This method is based on atomic feature optimization model, which uses the overall constraint information of feature model to divide the population into effective allocation and invalid allocation. Crossover and other genetic operations to accelerate to the optimal solution convergence.The experimental results show the effectiveness of the method.