Analysis and improvement of release readiness - A genetic optimization approach

Abstract

Context: Release readiness (RR) quantifies the status of a product release by aggregating a portfolio of release related measures. Early identification of factors responsible in improving RR (i.e. RR improvement factors) can help project managers to (re)allocate resources to improve processes to achieve higher level of RR score. Objective: This paper has two objectives: i) to identify time-dependent RR improvement (RRI) factor(s); and ii) to identify a budget allocation strategy for maximum improvement of RR score for the upcoming time interval. Method: RELREA is an existing approach that determines RR from aggregating the degree of satisfaction of a portfolio of release process, product, deployment and support related measures. The proposed method DAICO enhances RELREA by performing dynamic instead of static analysis. For that purpose, the RRI factors identification problem is formulated and solved as a genetic optimization problem. Subsequently, recommendations are generated for cost-optimized RR improvement. Results: We demonstrated the applicability of the DAICO method for release Publify 8.0 of an ongoing project Publify, hosted in GitHub OSS repository. Main contributions of this paper are: i) Formulating identification of RRI factors as an optimization problem, ii) Modeling and solving the problem using a GA, iii) Providing recommendations for cost-optimized RR improvement Conclusions: DAICO is a part of an ongoing effort to detect, and analyze RRI factors when achieving RR. This method is intended to detect RRI factors earlier and to guide the effort spent on improving RR.