A Comparative Study of the Particle Swarm Optimization and Genetic Algorithm for Software Evolution Process

Abstract

The software industry increasingly emphasizes the improvement of software quality. Modern development practices recognize that software should be built and maintained through a well-defined, structured process, consisting of organized and systematic steps for designing, developing, and evolving software systems. Petri net is applied for basic block of the software process. Path-based testing supports the software process by systematically identifying key execution paths to improve test planning and coverage. Using optimization techniques like Particle Swarm Optimization (PSO) helps efficiently generate test paths, enhancing the quality and effectiveness of process-level testing activities. Search-based optimization techniques improve decision making in the software process by transforming tasks such as cost estimation, release planning, and test generation into optimization problems. These methods efficiently explore large solution spaces to find near optimal solutions. Genetic algorithm (GA) is particularly effective due to their scalability and ability to handle multiple objectives, enhancing process efficiency, resource allocation, and software quality. In this paper, we present the development of a workflow software process modeled using Petri Nets. The approach helps in analyzing, controlling, and improving process execution within software development. We have proposed path testing technique applies PSO and GA to efficiently generate test paths that satisfy the all-uses criterion, improving planning and coverage in the software process. Comparative results show that PSO can achieve complete def–use path coverage with fewer generations than GA, enhancing testing efficiency and process effectiveness. The software industry increasingly emphasizes the improvement of software quality. Modern development practices recognize that software should be built and maintained through a well-defined, structured process, consisting of organized and systematic steps for designing, developing, and evolving software systems. Petri net is applied for basic block of the software process. Path-based testing supports the software process by systematically identifying key execution paths to improve test planning and coverage. Using optimization techniques like Particle Swarm Optimization (PSO) helps efficiently generate test paths, enhancing the quality and effectiveness of process-level testing activities. Search-based optimization techniques improve decision making in the software process by transforming tasks such as cost estimation, release planning, and test generation into optimization problems. These methods efficiently explore large solution spaces to find near optimal solutions. Genetic algorithm (GA) is particularly effective due to their scalability and ability to handle multiple objectives, enhancing process efficiency, resource allocation, and software quality. In this paper, we present the development of a workflow software process modeled using Petri Nets. The approach helps in analyzing, controlling, and improving process execution within software development. We have proposed path testing technique applies PSO and GA to efficiently generate test paths that satisfy the all-uses criterion, improving planning and coverage in the software process. Comparative results show that PSO can achieve complete def–use path coverage with fewer generations than GA, enhancing testing efficiency and process effectiveness.