Asymmetric distributed constraints satisfaction problems Conference Paper uri icon

abstract

  • Distributed constraint satisfaction problems (DisCSP s) with asymmetric constraints reflect the fact that agents may wish to retain their constraints private. Brito and Meseguer proposed a model for asymmetric constraints which they term Partially Known Constraints (PKC). In the PKC model each binary constraint is divided between the two constraining agents. In order to solve the resulting DisCSP with asymmetric constraints, a two phase asynchronous backtracking algorithm was proposed [BM03]. In the first phase an asynchronous backtracking algorithm is performed, in which only the constraints held by the lower priority agents are examined. When a solution is reached, a second phase is performed in which the consistency of the solution is checked again, according to the constraints held by the higher priority agents in each binary constraint. The present paper proposes a one phase …

publication date

  • October 1, 2005