Configuration strategies for evolutionary testing

  • Authors:

  • Xiaoyuan Xie;Baowen Xu;Changhai Nie;Liang Shi;Lei Xu

  • Affiliations:

  • Dept of Computer Science and Engineering, Southeast University, Nanjing, China and Jiangsu Institute of Software Quality, Nanjing, China;Dept of Computer Science and Eng., Southeast Univ., Nanjing, China and Jiangsu Inst. of Software Quality, Nanjing, China and Computer School, National Univ. of Defense Techn., Changsha, China and ...;Dept of Computer Science and Engineering, Southeast University, Nanjing, China and Jiangsu Institute of Software Quality, Nanjing, China;Dept of Computer Science and Engineering, Southeast University, Nanjing, China and Jiangsu Institute of Software Quality, Nanjing, China;Dept of Computer Science and Engineering, Southeast University, Nanjing, China and Jiangsu Institute of Software Quality, Nanjing, China

  • Venue:
  • COMPSAC-W'05 Proceedings of the 29th annual international conference on Computer software and applications conference
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Evolutionary Testing (ET) is a kind of efficient method of automatically test case generation. ET uses a kind of meta-heuristic search technique, the Genetic Algorithm, to convert the task of test case generation into an optimal problem. Nowadays, ET has been widely researched in many areas, especially in the GA configuration problem. In this paper, we suggest two strategies for the Genetic Algorithm configuration, to improve the performance of ET. One is Annealing Genetic Algorithm (AGA), which alters the mutation probability dynamically, and the other is Restricted Genetic Algorithm (RGA), which adds restrictions into fitness functions. The two strategies made ET hit the global optimal solution in fewer generations, and most offspring genes located in the legal domain.