Modeling and designing a robotic swarm: A quantum computing approach
Maria Mannone,Valeria Seidita,Antonio Chella
DOI: https://doi.org/10.1016/j.swevo.2023.101297
IF: 10.267
2023-06-01
Swarm and Evolutionary Computation
Abstract:Nature is a neverending source of inspiration for technology. Quantum physics suggests applications toward quantum computing. Swarms’ self-organization leads to robotic swarm developments. Here, quantum computing is applied to swarm robotics. We model local interactions with a quantum circuit, testing it on simulators and quantum computers. To relate local with global behavior, we develop a block matrix-based model. Diagonal sub-matrices contain information on single robots; off-diagonal sub-matrices are the pairwise interaction terms. Comparing different swarms means comparing different block matrices. Choosing initial values and computation rules for off-diagonal blocks (with a particular logic gate), different behaviors can be modeled. To show the global-behavior emergence, we propose a specific pairwise-interaction logic gate, embedding the corresponding quantum circuit in an ant-foraging-inspired algorithm. To implement a first application, we choose the foraging-ant behavior for its clarity and importance in nature, running experiments with toy swarms (3 and 10 robots). We consider ants’ individual and collective back-and-forth movements between the nest and the food source, analyzing the effect of entanglement. Our research can help shed light on quantum potentialities for swarms. The implications of our findings and results concern the future development of a decision-making system, based on the advantages of swarms and quantum computing. While an ant-foraging scenario is chosen as an example of application, our study is not focused on optimization. We present a new methodology, open to non-optimal solutions. Future developments can concern improvements toward optimization.
computer science, artificial intelligence, theory & methods