Automatic Generation Of Test Data Based On Particle Swarm Optimization

With the rapid development of computer technology, computer systems increased dramatically in size and complexity, the software development costs as a result of software failure caused economic losses are also increasing. Therefore, software quality has aroused great concern. Software testing is an important means to ensure software quality, more and more attentions. According to Boehm’s statistics, the total cost of software development, with the overhead on the test to account for 30% to 50% [1]. Therefore, software testing software has become an important part of the development process.According to different division of software testing method, and a different classification. According to whether the program execution can be divided into static testing and dynamic testing. Static testing, is not actually running the software under test, but only statically checking the program code, interface, or there may be errors in the document process. Dynamic testing is actually running the program under test, enter the appropriate test data, check the actual output and expected results are consistent with the process.According to whether to use software tools can be divided into manual and automated testing. Manual testing is a software test engineer, install and run the software under test through, according to the requirements of test documentation, run the test case, the results are correct observation software to run the process; automated testing is the use of a variety of automated software testing tools by running prior designed test scripts and test data file test test software to automatically generate test reports process.Thus, test data is an important part of software testing. The automatic generation of test data will effectively reduce the labor intensity of the test, saving the cost of software development.This year people try to use artificial intelligence technology in automatic test data generation, and has achieved initial results. Determine the characteristics of the test data generated in the field of artificial intelligence techniques will be promising. The basic idea of this method is to first randomly selected input data and run the program, and then based on the results in the core optimization algorithms to "evolve" a new input data continue to test until you find the correct solution, and therefore called heuristics. The current use of the basic genetic algorithm or improved genetic algorithm as a core optimization algorithm to automatically generate test data.The main work is summarized as follows :Systematic study of heuristic search algorithms in test data generation applications, summarizes the genetic algorithm, simulated annealing algorithm, PSO advantages and disadvantages.In the particle swarm algorithm based on the combination test procedures state diagram presents a method for generating test data. This method not only takes into account the type of data, the range is also considering using the state diagram between the input successively, causality, and the use of the characteristics of the particle swarm algorithm, in a multi- branching paths to simulate the user to select, in accordance with the emergence of successful path the frequency generated test data.Using the above study key algorithms and techniques, design automatic test data generation system, the use of technology-based Web to implement the system, using Jsp technology to achieve the system interface layer, using servlet implementation of user interaction logic control and data transfer, Java language algorithm design, the resulting data is stored in MySQL.