Comparative Evaluation of Intelligent Search Algorithms for Resource Allocation in Computing Environments

Abstract

Selecting a search algorithm for resource allocation requires more than comparing execution time. Completeness, optimality, memory demand, heuristic accuracy, adaptability, and implementation cost change with workload structure and operating constraints. This decision-oriented review synthesizes 289 publications on exact, heuristic, metaheuristic, and adaptive search methods. The evidence is organized around six dimensions: completeness and optimality, time and space complexity, heuristic quality, dynamic resource conditions, scalability and parallelism, and explainability and implementation cost. The synthesis shows that no algorithm dominates across sparse and dense state spaces, static and dynamic workloads, or exact and time-bounded objectives. Algorithm performance depends on the interaction between search guarantees, representation choices, heuristic information, computational resources, and stopping rules. The paper contributes a problem-to-algorithm decision framework, a comparative assessment matrix, and validation criteria that separate theoretical guarantees from empirical performance. The framework supports transparent algorithm selection and reproducible evaluation in computing environments.