Task Allocation Strategy Based on Variances in Bids for Large-Scale Multi-Agent Systems

Abstract : We propose a decentralized task allocation strategy by estimating the states of task loads in market-like negotiations based on an announcement-bid-award mechanism, such as contract net protocol (CNP), for an environment of large-scale multi-agent systems (LSMAS). CNP and their extensions are widely used in actual systems, but their characteristics in busy LSMAS are not well understood and thus we cannot use them lightly in larger application systems. We propose an award strategy in this paper that allows multiple bids by contractors but reduces the chances of simultaneous multiple awards to low-performance agents because this significantly degrades performance. We experimentally found that it could considerably improve overall efficiency.
Document type :
Conference papers
Complete list of metadatas

Cited literature [9 references]  Display  Hide  Download

https://hal.inria.fr/hal-01459603
Contributor : Hal Ifip <>
Submitted on : Tuesday, February 7, 2017 - 1:04:39 PM
Last modification on : Saturday, February 9, 2019 - 9:06:02 PM
Long-term archiving on : Monday, May 8, 2017 - 2:17:30 PM

File

978-3-642-41142-7_12_Chapter.p...
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

Citation

Toshiharu Sugawara. Task Allocation Strategy Based on Variances in Bids for Large-Scale Multi-Agent Systems. 9th Artificial Intelligence Applications and Innovations (AIAI), Sep 2013, Paphos, Greece. pp.110-120, ⟨10.1007/978-3-642-41142-7_12⟩. ⟨hal-01459603⟩

Share

Metrics

Record views

61

Files downloads

106