The swarm based algorithms can be modelled, under suitable assumptions, as equivalent dynamic circuits reproducing the cinematic characteristics of the trajectories followed by the swarm members. This can be made in terms of voltages measured at the terminal of capacitors and currents measured at the terminals of inductors. Through the use of swarm-circuits the role played by the parameters becomes clear since it is possible to apply the stability analysis of continuous systems. This allows us to govern the exploration and/or the exploitation properties of the system simply by tuning its parameters into the convergence range or vice versa. The presented circuital model has been tested on famous benchmarks for optimization and inverse problems. The obtained results show that the swarm circuits are capable to manage exploration as well exploitation and can be used for real-time optimizations such as navigation in unknown ambient of mobile robots and so on.
Laudani, A., Pulcini, G., RIGANTI FULGINEI, F., Salvini, A. (2014). Electric circuits performing the swarm optimization. INVERSE PROBLEMS IN SCIENCE & ENGINEERING, 22(7), 1109-1127 [10.1080/17415977.2013.854355].
Electric circuits performing the swarm optimization
LAUDANI, ANTONINO;RIGANTI FULGINEI, Francesco;SALVINI, Alessandro
2014-01-01
Abstract
The swarm based algorithms can be modelled, under suitable assumptions, as equivalent dynamic circuits reproducing the cinematic characteristics of the trajectories followed by the swarm members. This can be made in terms of voltages measured at the terminal of capacitors and currents measured at the terminals of inductors. Through the use of swarm-circuits the role played by the parameters becomes clear since it is possible to apply the stability analysis of continuous systems. This allows us to govern the exploration and/or the exploitation properties of the system simply by tuning its parameters into the convergence range or vice versa. The presented circuital model has been tested on famous benchmarks for optimization and inverse problems. The obtained results show that the swarm circuits are capable to manage exploration as well exploitation and can be used for real-time optimizations such as navigation in unknown ambient of mobile robots and so on.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.