Can we Knapsack Software Defect Prediction? Nokia 5G Case

As software products become larger and more complex, the test infrastructure needed for quality assurance grows similarly, causing a constant increase in operational and maintenance costs. Although rising in popularity, most Artificial Intelligence (AI) and Machine Learning (ML) Software Defect Prediction (SDP) solutions address singular test phases. In contrast, the need to address the whole Software Development Life Cycle (SDLC) is rarely explored. Therefore in this paper, we define the problem of extending the SDP concept to the entire SDLC, as this may be one of the significant next steps for the field. Furthermore, we explore the similarity between the defined challenge and the widely known Multidimensional Knapsack Problem (MKP). We use Nokia's 5G wireless technology test process to illustrate the proposed concept. Resulting comparison validates the applicability of MKP to optimize the overall test cycle, which can be similarly relevant to any large-scale industrial software develonment process.