The SWILT project is centered around managing collectives of cyber-physical systems (CPS) in industrial settings, such as groups of products or machinery. These systems are increasingly prevalent in various sectors, including the Internet of Things (IoT), smart mobility, energy distribution, and automated home and manufacturing systems. The complexity of coordinating these CPS swarms is considerable and challenging to direct and program. Inspiration from natural systems and their emergent collective behavior from individual interactions offers a way to manage complexity.

SWILT explores specific applications like production planning within industrial facilities and transportation logistics. Although linear optimization is the standard method in these scenarios, it is generally limited to a fraction of the operation due to growing complexity, rendering it insufficient for large-scale industrial scheduling tasks. The sizes of these real-world problems are very large, and fixed schedules or logistical plans cannot account for every variable, which is where the SWILT project innovates. It proposes the use of swarm intelligence algorithms that operate based on simple local rules for each agent. This approach allows for dynamic adjustment to changes in the system through local agent interactions.

Swarm Intelligence Layer to Control Autonomous Agents (SWILT)
Elnaz Khatmi [0000−0002−1016−3933],
Wilfried Elmenreich [0000−0001−6401−2658],
Kristina Wogatai [0000−0001−5246−0672],
Melanie Schranz [0000−0002−0714−6569],
Martina Umlauft, Walter Laure, and Andreas Wuttei

Link to Paper: https://ceur-ws.org/Vol-2405/16_paper.pdf