The Evolution of Metaheuristics: From Classical to Intelligent Hybrid Frameworks

Abstract

Metaheuristic algorithms have experienced unprecedented evolution over the past several decades, emerging as potent stochastic optimization tools across a wide spectrum of real-world applications. This article provides a comprehensive review of the evolution of metaheuristics, tracing their origins from classical trajectory-based and population-based approaches to the modern era characterized by intelligent hybrid frameworks that integrate machine learning, reinforcement learning, and adaptive parameter tuning. In the early stages, metaheuristics were mainly inspired by natural phenomena—from the cooling process in simulated annealing to the collective behaviors observable in swarm intelligence—thereby establishing a robust foundation for solving complex global optimization problems23. More recently, over 500 metaheuristic algorithms have been developed, with more than 350 emerging in the last decade alone, reflecting both the inventive spirit in algorithm design and an ongoing debate surrounding the novelty of seemingly similar methodologies1. An important contribution of this paper is the presentation of a new taxonomy based on the number of control parameters in metaheuristic algorithms, which helps to clarify the relationships among diverse algorithmic strategies1. Key aspects such as hybridization strategies and AI-driven adaptations are discussed in depth, showing how intelligent modifications can lead to significant performance improvements—for instance, reducing air traffic complexity by 92.8% within a hyper-heuristic framework leveraging reinforcement learning5. The evolution of metaheuristics is contextualized within their growing applications in engineering, healthcare, energy, telecommunications, and urban planning, underscoring their practical importance. Overall, this review not only synthesizes historical developments but also provides insights into current trends and emerging directions in metaheuristic research, with the goal of guiding both veteran researchers and newcomers in the field.