A Comprehensive Review of Ant Colony Optimization: Foundations, Variants, Applications, Challenges, and Future Research Directions

Abstract :

Ant Colony Optimization (ACO) is one of the most famous metaheuristics based on collective intelligence. It is a metaheuristic algorithm inspired from nature, by searching food process of ants associated with the pheromone utilization. The algorithm of ACO is a discrete metaheuristic that is appropriate to the problem solving combinatorial, continuous, multi-objective, dynamic and huge scale optimization problems. It is the purpose of this paper to express an analytical and coherent review of the biological and computational underpinnings, mathematical expression, mainspecies, parameters fine-tuning procedures, applications, the deficiencies and emerging research fields of ACO in the way of narrative-analytical review including up-to-date new articles until 2026.  Based on the review, the main strengths of ACO are the ability to model natural graph-based problems, distributed collective learning, the possibility of combining with local search, and flexibility in component design. In contrast, search stagnation, parameter sensitivity, complex computations in large dimensions, and difficulty in transferring settings between problems are some of the challenges of the proposed algorithm. Examples such as Ant Colony System, MAX–MIN Ant System, ACO for continuous problems, multi-objective methods, parallel versions, and guided learning algorithms have led to different challenges. Other recent developments involves a transition from predefined rules to automatic learning and tuning, neural networks and additive learning, parallel execution on GPU, optimization under uncertainty, as well as the development of reproducible and low-energy techniques. Last but not least, an applicative framework for the design and testing of ACO research is also presented.