Title: SmartSort: An Intelligent Framework for Optimizing Sorting Efficiency Using AI in Real-Time Systems
Authors: Khaleel Alnajjar and Samy S. Abu-Naser
Volume: 9
Issue: 8
Pages: 80-85
Publication Date: 2025/08/28
Abstract:
SmartSort is proposed as an adaptive hybrid sorting algorithm that leverages both QuickSort and HeapSort techniques, augmented by an AI-driven decision module to optimize performance under real-time constraints. This framework addresses the challenges of unpredictable input distributions and stringent timing requirements in real-time systems by combining the speed of QuickSort on average and the guaranteed worst-case behavior of HeapSort. SmartSort uses machine learning classifiers (e.g. decision trees, XGBoost or neural networks) to analyze data patterns - such as size, range, and entropy - and adapt its strategy (pivot selection or algorithm choice) on-the-fly[1][2]. We formally analyze SmartSort's time complexity and space usage, demonstrating average-case O(n log n) and worst-case O(n log n) behavior (due to the HeapSort fallback) while maintaining low overhead. The algorithm's best-, average-, and worst-case behaviors are discussed in detail, and we show that SmartSort can satisfy bounded-latency and predictability requirements typical of real-time systems[3][4]. Recent literature on AI-driven algorithm optimization and real-time scheduling is surveyed to support this design[5][6].