A flocking algorithm for multi-agent systems with connectivity preservation under hybrid metric-topological interactions

Autor: Zhigang Ren, Chenlong He, Zuren Feng
Jazyk: angličtina
Rok vydání: 2018
Předmět:
0209 industrial biotechnology
Correctness
Computer science
Velocity
lcsh:Medicine
02 engineering and technology
01 natural sciences
Upper and lower bounds
Systems Science
Topology
Database and Informatics Methods
020901 industrial engineering & automation
Agent-Based Modeling
lcsh:Science
Multidisciplinary
Behavior
Animal
Multi-agent system
Simulation and Modeling
Applied Mathematics
Physics
Classical Mechanics
Eukaryota
Graph
Algebraic Topology
Physical Sciences
Vertebrates
Heuristic Alignment Procedure
Graph (abstract data type)
Sequence Analysis
Algorithms
Research Article
Computer and Information Sciences
Bioinformatics
Algebraic topology
Research and Analysis Methods
Birds
Motion
0103 physical sciences
Animals
Computer Simulation
010306 general physics
Flocking (behavior)
lcsh:R
Organisms
Biology and Life Sciences
Models
Theoretical
Computer Science::Multiagent Systems
Nonlinear system
Algebra
Flight
Animal
Amniotes
lcsh:Q
Mathematics
Zdroj: PLoS ONE
PLoS ONE, Vol 13, Iss 2, p e0192987 (2018)
ISSN: 1932-6203
Popis: In this paper, we propose a connectivity-preserving flocking algorithm for multi-agent systems in which the neighbor set of each agent is determined by the hybrid metric-topological distance so that the interaction topology can be represented as the range-limited Delaunay graph, which combines the properties of the commonly used disk graph and Delaunay graph. As a result, the proposed flocking algorithm has the following advantages over the existing ones. First, range-limited Delaunay graph is sparser than the disk graph so that the information exchange among agents is reduced significantly. Second, some links irrelevant to the connectivity can be dynamically deleted during the evolution of the system. Thus, the proposed flocking algorithm is more flexible than existing algorithms, where links are not allowed to be disconnected once they are created. Finally, the multi-agent system spontaneously generates a regular quasi-lattice formation without imposing the constraint on the ratio of the sensing range of the agent to the desired distance between two adjacent agents. With the interaction topology induced by the hybrid distance, the proposed flocking algorithm can still be implemented in a distributed manner. We prove that the proposed flocking algorithm can steer the multi-agent system to a stable flocking motion, provided the initial interaction topology of multi-agent systems is connected and the hysteresis in link addition is smaller than a derived upper bound. The correctness and effectiveness of the proposed algorithm are verified by extensive numerical simulations, where the flocking algorithms based on the disk and Delaunay graph are compared.
Databáze: OpenAIRE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje