This special issue of the Swarm Intelligence journal is dedicated to the publication of extended versions of some of the best papers presented at ANTS 2022, Thirteenth International Conference on Swarm Intelligence, which took place in Málaga, Spain, November 2–4, 2022.

ANTS is the first and most established conference series dedicated to the dissemination of swarm intelligence research. Its first edition took place in 1998, at the Université Libre de Bruxelles, Brussels, Belgium. Since then, it has been held every other year and, since 2010 (i.e., starting with the seventh edition of the conference), the authors of full conference papers have been invited to submit extended versions for possible inclusion in a dedicated special issue of the Swarm Intelligence journal.

Six papers from the 2022 edition of the ANTS conference were accepted for publication in the special issue after at least two rounds of reviews with comments by at least three referees. The papers co-authored by a guest editor of this special issue have been managed anonymously by one of the other guest editors. In the following, each paper is briefly introduced by the guest editor who managed it.

With the paper Decomposition and Merging Co-operative Particle Swarm Optimization with Random Grouping for Large-Scale Optimization Problems, Alanna McNulty, Beatrice Ombuki-Berman, and Andries Engelbrecht work on the challenge of large-scale optimization problems. Starting from Cooperative Particle Swarm Optimization (CPSO), they introduce two novel methods in the context of decomposing and merging CPSO. They incorporate random grouping of decision variables and improve variable dependency exploration. Their experiments demonstrate that these new approaches outperform or are competitive with established PSO methods. The study highlights the importance of exploration in the context of random grouping to effectively solve large problem instances.

In the paper On the Automatic Design of Multi-Objective Particle Swarm Optimizers: Experimentation and Analysis, Antonio J. Nebro, Manuel López-Ibáñez, José García-Nieto, and Carlos A. Coello Coello address the challenge of designing effective multi-objective particle swarm optimizers (MOPSOs). The authors present AutoMOPSO, a flexible algorithmic template for designing MOPSOs that is capable of instantiating thousands of potential designs. By implementing AutoMOPSO within the jMetal framework and applying the automatic algorithm configuration tool irace, the authors show that AutoMOPSO can be used to automatically generate high-performance MOPSO implementations that outperform state-of-the-art algorithms. The authors conduct an in-depth analysis and address research questions related to the performance, accuracy, influential design parameters, and generalization capabilities of AutoMOPSO. The paper contributes an automatic design methodology for MOPSOs that provides insights into key design choices and parameter settings, with potential applications to real-world problem scenarios.

The paper Predictive Search Model of Flocking for Quadcopter Swarm in the Presence of Static and Dynamic Obstacles, by Giray Önür, Ali Emre Turgut, and Erol Şahin, deals with a key problem in bringing swarm robotics outside of laboratory settings: coordinated motion with large-scale swarms of robots, where the robots are incapable of global localization. The main idea is to equip the robots with the ability to predict the future states of the swarm, expressed as a state tree. The robots use the state tree to search for the best target state to achieve next. The authors conducted an experimental campaign that involved simulations and real drones, and showed that the proposed approach allows large swarms to navigate safely through both static and dynamic obstacles.

With the paper Improved Decentralized Cooperative Multi-Agent Path Finding for Robots with Limited Communication, Abderraouf Maoudj and Anders Lyhne Christensen tackle the challenges of multi-agent path finding (MAPF) for robots with limited communication capabilities. They present an enhanced decentralized method that enables robots to coordinate using only local communication, effectively planning and replanning paths to avoid collisions, deadlocks, and livelocks. Through simulations with up to 1,000 robots, their method demonstrates scalability and efficiency. Their approach competes with traditional centralized approaches and offers a robust solution for real-world applications where global communication is constrained or impractical.

In the paper Many-option Collective Decision Making: Discrete Collective Estimation in Large Decision Spaces by Qihao Shan and Sanaz Mostaghim, the authors compare collective decision-making strategies in scenarios where the number of options is high, even higher than the number of agents. In their in-depth simulation study with varying decision spaces, the authors highlight that two discrete decision-making strategies (iterative ranked voting and Bayesian belief fusion) can outperform—in terms of error and convergence time—a commonly used continuous consensus mechanism (linear consensus protocol) in many cases. Overall, the study shows that the discretization of a consensus problem can be advantageous for reaching fast and exact consensus in distributed systems, provided that the discretization is not overly fine.

Aadesh Neupane and Michael A. Goodrich, in Resilient Swarm Behaviors via Online Evolution and Behavior Fusion, propose an enhanced approach to the well-known grammatical evolution approach to automatic design of behavioral logic. The authors’ observation is that traditional approaches to grammatical evolution adapt slowly when behaviors need to be resilient to perturbations. They suggest that this is due to traditional algorithms only relying on vertical, that is, parent-to-child, genetic variation. They propose a novel algorithm, called BeTr-GEESE, which allows also for “lateral” gene transfer, where genes are transferred across individuals within the same generation. The approach is validated through experiments involving foraging and nest maintenance. In addition, the authors also present a variant of their algorithm, called Multi-GEESE, that addresses the challenge of staying resilient in dynamic environments even after evolution is stopped.

The high quality of the six papers contained in this special issue is the result of the collaboration of a large number of people: the authors, who submitted their best work to the journal; the referees, who helped in the selection of the published papers; and finally, the many people at Springer who assisted us in the production phase. We thank them all for their help to make this special issue possible.