Affordable Access

pyDCOP, a DCOP library for IoT and dynamic systems

Authors
  • Rust, Pierre
  • Picard, Gauthier
  • Ramparany, Fano
Publication Date
May 14, 2019
Source
HAL-EMSE
Keywords
Language
English
License
Unknown
External links

Abstract

This demonstration illustrates the newly developed Python-based framework, pyDCOP, which implements several state-of-the-art distributed constraint reasoning solution methods, provides utilities to deploy them over distributed infrastructures and also equip the system with resilience capabilities.The idea behind pyDCOP is to distribute agents over an Internet-of-Things infrastructure (e.g. Rapsberry Pis) to install collective decisions, as to implement Ambient Intelligence or Smart Home scenarios. Scenarios are modeled in a dedicated format, translated in a distributed constraint optimization or satisfaction problem, then pushed to the devices which coordinate using chosen protocols as to self-configure in a decentralized manner. Besides configuring the system in an optimal manner, it also provides a resilience framework, which equips the system with adaptation capabilities against unpredictable device removals. This mechanism is based on decision replication and a lightweight DCOP-based reparation mechanism.

Report this publication

Statistics

Seen <100 times