Distributed Search by Constrained Agents

  • Amnon Meisels
Part of the Studies in Computational Intelligence book series (SCI, volume 382)


The investigation of Distributed Constraints Satisfaction Problems (DCSPs) has started a little more than a decade ago. It focuses on constraints satisfaction problems (CSPs) that are distributed among multiple agents. Imagine a large University that includes many departments. The weekly schedule of classes is generated by each department, scheduling its classes and teachers for the whole semester. A weekly schedule is a typical constraint satisfaction problem. Class meetings are variables and the time-slots of the week are the domain of values that have to be assigned to classes in order to generate a schedule. The fundamental constraints of timetabling require that two classes taught by the same teacher have to be assigned different time-slots. Another common constraint is to require that two meetings of the same class will be assigned to different days of the week. In the University as a whole, the departments can be thought of as agents that generate their departmental weekly schedules. The weekly schedules of different departments are constrained by the fact that there are students that select classes from these departments. This generates constraints between departments and the generic scenario of a distributed CSP. Agents own parts of the global problem (e.g. departmental schedules) and cooperate in search for a global solution in which the constraints between departments are satisfied. In order to solve such a distributed problem, all agents must cooperate in a global search process. Search algorithms for a distributed problem operate by agents performing assignments to their variables and exchanging messages in order to check their assignments against those of constraining agents.


Constraint Satisfaction Problem Weekly Schedule Class Meeting Global Search Algorithm Common Constraint 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Gershman, A., Meisels, A., Zivan, R.: Asynchronous forward bounding. J. of Artificial Intelligence Research 34, 25–46 (2009)MathSciNetGoogle Scholar
  2. 2.
    Modi, P.J., Shen, W., Tambe, M., Yokoo, M.: ADOPT: Asynchronous distributed constraints optimization with quality guarantees. Artificial Intelligence 161(1-2), 149–180 (2005)MathSciNetzbMATHCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Amnon Meisels
    • 1
  1. 1.Ben Gurion University of the NegevBe’er ShevaIsrael

Personalised recommendations