Difference between revisions of "Self-organizing synchronization"
From Self-Organization Wiki
| (One intermediate revision by the same user not shown) | |||
| Line 3: | Line 3: | ||
= Application areas= | = Application areas= | ||
* Smart grids | * Smart grids | ||
| − | The goal is to balance the load across the network. | + | ** The goal is to balance the load across the network. |
| − | A phase-locked system is sufficient (not necessarily in-phase). | + | ** A phase-locked system is sufficient (not necessarily in-phase). |
* Wireless systems | * Wireless systems | ||
| − | Synchronization within 1% of the slot duration is sufficient. | + | ** Synchronization within 1% of the slot duration is sufficient. |
| − | = Main issues= | + | = Main general issues= |
* Robustness to faulty nodes | * Robustness to faulty nodes | ||
| − | Does the system return in place after a node fails/does not follow local rules? | + | ** Does the system return in place after a node fails/does not follow local rules? |
* Fault tolerance | * Fault tolerance | ||
* Scalability | * Scalability | ||
| − | Hierarchical synchronization, i.e. clustering, can be applied on top of SO Sync to limit scalability issues. | + | ** Hierarchical synchronization, i.e. clustering, can be applied on top of SO Sync to limit scalability issues. |
* Implementation | * Implementation | ||
* Synchronization as a primal form of coordination | * Synchronization as a primal form of coordination | ||
| + | * Optimum solution | ||
| + | ** Finding a method that is most robust and scalable in meshed networks | ||
= Open issues= | = Open issues= | ||
* Better understanding of inhibitory behavior in meshed networks | * Better understanding of inhibitory behavior in meshed networks | ||
| + | ** Why is inhibitory coupling worse in sparse networks? | ||
* Dynamic networks | * Dynamic networks | ||
| + | ** Hassler and Toroczkai, 2005 (distributing locals onto computers s.t. they finish computing at the same time) | ||
| + | ** Stochastic oscillators | ||
* MEMFIS | * MEMFIS | ||
| − | ** Mathematical proof | + | ** Mathematical proof for excitatory coupling in meshed networks. First clue in Restrepo and Ott paper, Phys. Rev. E |
** Time to synchrony increases with the network diameter | ** Time to synchrony increases with the network diameter | ||
* Discrete Kuramoto model | * Discrete Kuramoto model | ||
Latest revision as of 13:37, 29 July 2010
Contents
Goal of the session: identifying research issues
Application areas
- Smart grids
- The goal is to balance the load across the network.
- A phase-locked system is sufficient (not necessarily in-phase).
- Wireless systems
- Synchronization within 1% of the slot duration is sufficient.
Main general issues
- Robustness to faulty nodes
- Does the system return in place after a node fails/does not follow local rules?
- Fault tolerance
- Scalability
- Hierarchical synchronization, i.e. clustering, can be applied on top of SO Sync to limit scalability issues.
- Implementation
- Synchronization as a primal form of coordination
- Optimum solution
- Finding a method that is most robust and scalable in meshed networks
Open issues
- Better understanding of inhibitory behavior in meshed networks
- Why is inhibitory coupling worse in sparse networks?
- Dynamic networks
- Hassler and Toroczkai, 2005 (distributing locals onto computers s.t. they finish computing at the same time)
- Stochastic oscillators
- MEMFIS
- Mathematical proof for excitatory coupling in meshed networks. First clue in Restrepo and Ott paper, Phys. Rev. E
- Time to synchrony increases with the network diameter
- Discrete Kuramoto model
