•  
  •  
 

Abstract

The software Requirements Prioritization (RP) process is essential for producing a successful software project. Requirements are interdependent in software projects, so handling their dependency during the RP process is mandatory. Many researchers have shown that requirements dependency is challenging for large-scale systems. Extracting requirements dependency is difficult since requirements are documented in natural language. Improper handling of dependencies among requirements while prioritization can cause inaccurate prioritization results and deadlocks, which cause project delays, rework, and redesign. Many techniques have been introduced to automate the dependency analysis process among software requirements, including artificial intelligence (AI) and other logic-based methods such as fuzzy logic and fuzzy graphs. However, these techniques have significant variations in detecting hidden dependencies, scalability, accuracy, adaptability, and transparency, raising challenges for their applicability in real-world projects. This paper conducts a systematic literature review of software requirements dependency analysis techniques by selecting 34 primary articles from different sources from 2019 to 2025 with specific selection criteria. It provides a comparative analysis among 28 existing dependency analysis techniques by comparing them through several key attributes. The research reveals critical gaps affecting their suitability in real-world projects. It highlights a tradeoff between accuracy and explainability in the reviewed techniques. It also shows limitations in their ability to handle requirements dependency changes, scalability, and generalizability, offering a clear foundation for future research direction and emphasizing the need to develop approaches that combine all these attributes to achieve a high-quality dependency analysis process.

Share

COinS