Optimal Task Allocation in Sensor and Actuator Networks
2015 (English)In: Proc. 11th Swedish National Computer Networking Workshop, 2015Conference paper (Refereed)
Sensor and actuator network macroprogramming techniques that use task graphs are promising options for high-level application development in this domain. However, setting up a multitude of application-level tasks in the network and subsequently keeping track of them is a nontrivial undertaking. Therefore, tool support for automated task allocation is required. Such a support is typically implemented by providing an objective function that evaluates the cost of a task mapping, and a search algorithm that attempts to minimize this function.The current algorithms for task allocation in sensor and actuator networks either do not guarantee optimal results, or are conceptually tied to a single specific objective function. Our work contributes to this state-of-art in two aspects: firstly, by finding a model that can easily accommodate several different objective functions, and secondly, by finding a search strategy that makes it feasible to allocate realistic task graphs even in large networks.
Place, publisher, year, edition, pages
IdentifiersURN: urn:nbn:se:uu:diva-256753OAI: oai:DiVA.org:uu-256753DiVA: diva2:826810
SNCNW 2015, May 28–29, Karlstad, Sweden
FunderSwedish Foundation for Strategic Research , RIT08-0065