Automatic Generation of Test Cases Based On Multi-population Genetic Algorithm

Content Provider	Semantic Scholar
Author	Zhang, Na Wu, Biao Bao, Xiaoan
Copyright Year	2015
Abstract	The design of automatic generation technology of test case is an important part of the software test automation implementation, having an important guiding role in testing of late work, which is the fundamental guarantee to improve the reliability of software. In this paper, considering the lack of adequacy of control flow testing, using the data flow testing as the test adequacy criteria, and then on the basis of the single population genetic algorithm search efficiency is not high, combining with previous methods on the improvement of the genetic algorithm, introducing the concept of multi-population, and then designs a kind of improved parallel evolutionary algorithm (IPEA) based on multipopulation is used to automatically generate test cases. The algorithm defined the concept of external pressure which as the degree of competition between individuals. Fully considering the influence of coverage, branch condition and degree of competition between individual species of three aspects, and give different weights, we design a fitness function to evaluate the merits of the individual species. Experiments show that the IPEA has obviously improved in convergence speed, search time, coverage, scale of the test cases on key performance than the single population genetic algorithm and random search algorithm.
File Format	PDF HTM / HTML
Alternate Webpage(s)	http://www.sersc.org/journals/IJMUE/vol10_no6_2015/11.pdf
Language	English
Access Restriction	Open
Subject Keyword	Assistive technology Control flow Dataflow Evaluation function Evolutionary algorithm Experiment Fitness function Genetic algorithm Genetic operator Numerous Random search Search algorithm Simulation Software bug Software testing Test automation Test case Weight
Content Type	Text
Resource Type	Article