摘要
总结了测试数据自动生成的通用模型,把模拟退火遗传算法(SAGA)引入模型中作为核心搜索方法,避免了单一遗传算法易陷入局部最优的问题。在算法中,通过实施模拟退火筛选、改进加速适应度函数和控制温度迭代方式达到了模拟退火算法(SA)与遗传算法(GA)的有机融合,并优化了SA的邻域构造方式,加入了遗传算法的最优保留策略,使得算法在收敛效率等方面更加有效。试验证明,该方法是有效的测试数据自动生成方法。
The general model of dynamic test data generation is discussed in this paper. Simulated annealing genetic algorithm（SAGA） is utilized in this model as the main searching method for solving the problem of trapping in local optimum. In this algorithm, the implementation of simulated annealing filtration, the improvement of accelerated-fitness function and the control of temperature＇s iterative style construction combine simulated annealing（SA） and genetic algorithm（GA） together. Furthermore, the neighbor construction of SA is optimized and elitism reservation method of GA is applied to improve convergence efficiency. Finally, some experiment results are presented which show that SAGA is more efficient than other methods.
出处
《信息与电子工程》
2005年第1期40-43,共4页
information and electronic engineering
