•  
  •  
 

Abstract

Code clone detection is a critical task in software engineering, directly influencing code maintainability, quality, and technical debt management. Despite advancements in both static analysis techniques and AI-driven approaches, existing solutions remain fragmented, either relying on deterministic syntactic analysis for syntactic clone detection or employing machine learning, deep learning, and LLM-based methods for semantic clone detection. This separation leads to inefficiencies, scalability challenges, and missed optimization opportunities. This paper introduces a Hybrid Intelligent Clone Detection Architecture (HICDA), a context-driven delegation framework that routes clone detection tasks based on syntactic or semantic characteristics. Central to this architecture is the conceptualization of a Context-Driven Delegation Controller, which dynamically assigns detection tasks based on the syntactic or semantic complexity of cloned code fragments, optimizing resource utilization by integrating deterministic precision with semantic reasoning. HICDA is embedded within a comprehensive code clone management framework, incorporating automated refactoring and post-detection code quality evaluation, seamlessly integrated into real-time development workflows. Theoretical justification is provided through efficiency models, cognitive delegation analogies, and alignment with modern software architecture principles. A case study demonstrates the framework’s practical viability, highlighting reductions in computational overhead without compromising detection accuracy. This contribution establishes a foundational model for hybrid AI-augmented static analysis, offering a scalable and generalizable blueprint for future intelligent software engineering tools.

Keywords

Code clone detection, Hybrid architecture, Static analysis, Large language models, Software maintainability, Semantic reasoning, Refactoring

الكلمات المفتاحية

AI generated arabic text detectors, DNN, ChatGPT, NLP, LSTM, Word embedding, SSA, CNN

Article Type

Article

First Page

44

Last Page

50

Publication Date

12-31-2025

Download

Share

COinS