A greedy particle swarm optimization (GPSO) algorithm for testing real-world smart card applications

作     者：Allawi, Hamzeh M. Al Manaseer, Waref Al Shraideh, Mohammad 

作者机构：Univ Jordan Dept Comp Sci Amman Jordan 

出 版 物：《INTERNATIONAL JOURNAL ON SOFTWARE TOOLS FOR TECHNOLOGY TRANSFER》 (国际技术传输软件工具杂志)

年 卷 期：2020年第22卷第2期

页      面：183-194页

核心收录：

学科分类：08[工学] 0835[工学-软件工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：Software testing Java smart card Particle swarm optimization algorithm Greedy algorithm Genetic algorithm 

摘      要：Software testing continues to be regarded as a necessary and critical step in the software development life cycle. Among the multitudes of existing techniques, particle swarm optimization (PSO) algorithm, in particular, has shown superior merits for automatically generating software test cases for its easy implementation and for relying on fewer parameters that require tuning. Hence, several state-of-the-art PSO-based algorithms have been successfully used as a test data generator. On the other hand, greedy-based algorithms, which are commonly used to solve complex and multi-step combinatorial problems, are starting to gain momentum as a solution for the complexity problem of software testing. Greedy algorithms favored over other techniques when the solution of the problem is guaranteed to be near-optimal. As a result, the utilization of both greedy and PSO algorithms in a single solution for automatically generating test data represents a strong candidate if designed carefully. In this paper, we propose a novel hybrid greedy and PSO algorithm (GPSO) that jointly guarantees the effectiveness and close to optimality results for generating a minimum number of test data. Compared with the widely employed genetic algorithm (GA), our proposed GPSO outperforms the GA in terms of the average number of iterations, execution time, and coverage percentage. Experimental trials with six different typical Java card applications show that the use of the proposed GPSO as a test data generator results in an outstanding performance.