Enhanced Software Performance Testing for Big Data Platforms Using Clock-Controlled Computation Tree Logic with Particle Swarm and Genetic Optimization

Abstract

This research aims to solve the problems of testing inefficiency and lack of accuracy in software testing, and proposes a software performance testing system for big data platforms based on the clock-controlled computational tree logic method. The particle swarm algorithm finds the optimal solution through the movement and mutual cooperation of particles in the search space. Genetic algorithm evolves the population through selection, crossover, and mutation operations, ultimately finding the optimal solution. Secondly, long short-term memory networks and linear autoregressive models also have advantages in software testing, which can improve the effectiveness and efficiency of software testing through reasonable selection and combined use. The new algorithm utilizes the ability of PSO and GA algorithms to search for optimal solutions through particle motion and group cooperation in the search space, in order to determine key moment parameters and other relevant information in software testing systems. The research uses the algorithmic logic of the particle swarm algorithm and the genetic algorithm to confirm the moment parameters and other information of the software testing system. At the same time, an algorithmic model research on the joint coverage and the use of the value of the system, and finally makes use of the big data platform to analyze the research system. The specific indicators used in the study include 100% test case coverage, as well as the functional coverage of genetic algorithms and particle swarm optimization algorithms. The innovative combination of CCTL method and optimization algorithm in the research has improved the accuracy and stability of software testing. CCTL is an extended computational tree logic that introduces the concept of time, allowing testers to explicitly specify time constraints in software testing, thereby more accurately simulating real-world scenarios. The research results show that using the system to test software can achieve a coverage rate of 100% for its component use cases, while the functional coverage rates of genetic algorithm and particle swarm algorithm reach 90.36% and 91.32%, respectively. The accuracy of software testing research methods is 5% and 6% higher than that of LSTM and LAR methods. When the moment range of the particle parameter position information of the model is [150 ms, 250 ms], the maximum value of the target parameter velocity is 80 m/s and the minimum value is 0 m/s. The maximum value of the target azimuth velocity is 20 rad/s, and the minimum value is 0 rad/s. The system is able to determine the various parameters of the software, and at the same time in the software test results on the test results are normal, fault analysis can be completed normally, the performance of the algorithm is also superior to other algorithm models such as LSTM and LAR, and the study of the use of algorithms with a higher degree of stability. It can be seen that the system and methodology used in this research is superior to traditional methods and the test results of software testing have improved. This study provides a new research direction for platform software afterwards.