<![CDATA[Although data retrieval is a fundamental task in any computational system, search algorithms are often used statically and ignore runtime dataset properties. This frequently results in wasted computing power in structured, semi-structured, and unstructured data environments. This study presents a lightweight dynamic search algorithm selection framework called the Intelligent Search Adaptation Mechanism (ISAM), which operates at runtime and uses the cardinality, sortedness, and uniformity of the dataset's value distribution as criteria for selecting the optimal search algorithm. ISAM is an adaptive dispatch mechanism for selecting intelligent algorithms, unlike traditional algorithms.ISAM is an intelligent algorithm selection unified dispatch mechanism. ISAM has been tested with synthetic data sets of 1,000 to 1,000,000 elements implemented in C++. For unsorted datasets, the overhead incurred by ISAM is negligible, and for sorted datasets, it can reduce query latency by up to 73.1× over a non-adaptive baseline of Sequential Search. In fact, the retrieval performance is found to be close to the theoretical complexity of the chosen search algorithm, through scalability analysis. The results show how intelligent runtime search orchestration works well in a heterogeneous data environment.]]>
Copyrights © 2026
