Bio inspired framework for optimized test case generation and prioritization using machine learning

Abstract

Software Testing (ST) plays a dominant role in ensuring quality and security with the rapid advancement of Internet services and applications. Thus, in ST, Test Case Generation (TCG) is a crucial activity that involves creating specific Test Cases (TCs) or scenarios to verify and validate different aspects of a Software System (SS). In addition, the TCG process deals with the generation of a set of testing conditions, which could be utilized to validate the application s adequacy. However, the major concern in the research domain is to generate optimized test data that covers the entire critical path. Many researchers and practitioners have implemented numerous frameworks related to optimized TCG in ST over the past decade. In recent times, meta-heuristic algorithms have been utilized for generating TCs for multiple Path Coverage (PC) in one run. Numerous nature-inspired meta heuristic optimization techniques that are applied in TCG encompass Genetic Algorithm (GA), Particle Swarm Optimization (PSO), cuckoo search, Ant Colony Optimization (ACO), and so on. In addition, the Fitness Function (FF) guides those techniques to determine the search results quality. However, the conventional methodologies had some drawbacks, such as a lack of domain knowledge, lack of TC reviews, repetitive TCs, and incomplete requirements. This thesis implements robust architecture for promising TCG and optimal Test Case Prioritization (TCP) by addressing those research gaps. The proposed frameworks are generalized well enough to render the most informative TCs, upgrading the software quality and requirements. In this thesis, the four effective methodologies are encompassed for TCG and TCP. newline