Managing Software Testing Technical Debt Using Evolutionary Algorithms

Abstract

Technical debt (TD) happens when project teams carry out technical decisions in favor of a short-term goal(s) in their projects, whether deliberately or unknowingly. TD must be properly managed to guarantee that its negative implications do not outweigh its advantages. A lot of research has been conducted to show that TD has evolved into a common problem with considerable financial burden. Test technical debt is the technical debt aspect of testing (or test debt). Test debt is a relatively new concept that has piqued the curiosity of the software industry in recent years. In this article, we assume that the organization selects the testing artifacts at the start of every sprint. Implementing the latest features in consideration of expected business value and repaying technical debt are among candidate tasks in terms of the testing process (test cases increments). To gain the maximum benefit for the organization in terms of software testing optimization, there is a need to select the artifacts (i.e., test cases) with maximum feature coverage within the available resources. The management of testing optimization for large projects is complicated and can also be treated as a multi-objective problem that entails a trade-off between the agile software’s short-term and long-term value. In this article, we implement a multi-objective indicator-based evolutionary algorithm (IBEA) for fixing such optimization issues. The capability of the algorithm is evidenced by adding it to a real case study of a university registration process.