Multi-agent collaborative service and distributed problem solving

Jiming Liu*, Xiaolong Jin, Yi Tang

*Corresponding author for this work

Research output: Contribution to journalJournal articlepeer-review

13 Citations (Scopus)


In this paper, we show that collaborative services can be formulated into distributed constraint satisfaction problems. We introduce the notion of multi-agent collaborative service (MACS), and employ a distributed discrete Lagrange multipliers (DDLM) method to automatically handle an MACS task. The DDLM method is based on a distributed multi-agent system. The behaviors of agents are guided by predefined DDLM rules. In order to make it more efficient in achieving a solution state, we incorporate strategies for tuning the Lagrange multipliers. We validate the effectiveness of the DDLM method with benchmark SAT problems. Furthermore, we provide the mathematical properties of DDLM and present the corresponding DDLM algorithms.

Original languageEnglish
Pages (from-to)191-206
Number of pages16
JournalCognitive Systems Research
Issue number3
Publication statusPublished - Sept 2004

Scopus Subject Areas

  • Experimental and Cognitive Psychology
  • Cognitive Neuroscience
  • Artificial Intelligence

User-Defined Keywords

  • Constraint satisfaction problems
  • Distributed problem solving
  • Lagrange multipliers
  • Multi-agent collaborative services
  • Satisfiability problems


Dive into the research topics of 'Multi-agent collaborative service and distributed problem solving'. Together they form a unique fingerprint.

Cite this