1 / 10

Compilation of Distributed Constraint Satisfaction Problems

Compilation of Distributed Constraint Satisfaction Problems. Henry Kautz AT&T Labs / University of Washington. Ideas. 1. Can pre-compile constraint satisfaction problems so that solutions can be quickly found at run-time.

eryk
Download Presentation

Compilation of Distributed Constraint Satisfaction Problems

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Compilation of Distributed Constraint Satisfaction Problems • Henry Kautz • AT&T Labs / University of Washington

  2. Ideas • 1. Can pre-compile constraint satisfaction problems so that solutions can be quickly found at run-time. • 2. Can extend approach to distributed problems by employing representations based on Quantified Boolean formulas

  3. 1. Compilation • Goal: shift online costs offline • Simple approach: solve a particular problem instance in advance, cache solution • More general approach: compile a (small) specification of a (large) set of instances, so that each instance becomes easy to solve(Selman & Kautz 1996) • Specification: formula + “clamped” variables • Each setting of clamped variables gives a different instance • Compilation based on identifying backbone: variables whose values are fixed(Monasson, Zecchina, Kirkpatrick, Selman, & Troyansky 1999)

  4. Backbone Size

  5. Effectiveness of Compilation

  6. Properties of Compiled Problems

  7. Distributed CSPs v3 v4 v5 v1 v6 v2

  8. Quantified Boolean Formulas • SAT = Prototypical problem for centralized constraint satisfaction • QSAT (Quantified Boolean Satisfiability) = Prototypical problem for distributed CSP • Blue solver: Find settings for v1, v2 that are consistent with any possible choices made by the other solvers v1,v2v3,v4,v5,v6.P(v3,v4,v5,v6) B(v1,v2,v3,v4,v5,v6)

  9. Phase Transitions in QSAT • QSAT recently shown to exhibit same phase transition phenomena as SAT(Gent & Walsh 1999) • Hardness also depends on pattern of quantifiers • In Distributed CSP, models pattern of communication between solvers • Suggests that backbones can be found in QSAT problems as well!

  10. Research Proposal • Algorithms for compilation of QSAT problem classes by identification of backbone variables • Goal: fast, robust solution of distributed constraint satisfaction problems • Applications: • Logistics resource allocation • Real-time network reconfiguration • Distributed constraint-based planning

More Related