1 / 12

Reducing WCNF-3SAT to WCNF-2SAT - (among other things)

Reducing WCNF-3SAT to WCNF-2SAT - (among other things). A presentation of results in: Rod G. Downey and Michael R. Fellows Fixed-parameter tractability and completeness II: on completeness for W[1], Theoretical Computer Science 141 , pp. 109-131, (1995). Presentation by Nick Neumann.

maire
Download Presentation

Reducing WCNF-3SAT to WCNF-2SAT - (among other things)

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. Reducing WCNF-3SAT toWCNF-2SAT-(among other things) A presentation of results in: Rod G. Downey and Michael R. Fellows Fixed-parameter tractability and completeness II: on completeness for W[1], Theoretical Computer Science 141, pp. 109-131, (1995). Presentation by Nick Neumann

  2. Outline of talk • Summary of paper • Necessary definitions • Lemma for proving reduction • Result yielding reduction • Other results in paper/implications • Conclusion

  3. Summary of paper • Defines W[t] differently than in class(makes W[1] much “bigger”) • Shows collapse of W[1] to WCNF-2SAT- . • In process, shows WCNF-3SAT <=fpt WCNF-2SAT- • Result immediately yields several problems to be W[1]-complete

  4. Necessary definitions • Small gates – bounded fan-in • Large gates – unrestricted fan-in • Depth (of circuit) – max # gates on path from literal to output in circuit • Weft (of circuit) – max # LARGE gates on path from literal to output in circuit

  5. More definitions • LF(t,h) = language of circuits of weft <= t, depth <= h, satisfied by weight k assignment • W[t]: LÎW[t] if L <=fpt LF(t,h) for some h • W[1,s]: Weft 1, depth 2 circuits with fan- in for small gates bounded by s • W[t] bounds weft at t;W[1,s] bounds fan-in at s, fixes weft at 1, depth at 2

  6. Result yielding reduction • Paper proves W[1]=W[1,2] • Does this as follows: • Proves W[1,s]=antimonotone W[1,s] • Proves W[1]=Ès=1..∞ W[1,s] • Proves antimonotone W[1,s]=W[1,2]

  7. Lemma for proving reduction • W[1,s]=antimonotone W[1,s] for s>=2 • Proof strategy: • Takes a problem in antimonotone W[1,s] and shows it is hard for W[1,s] • Problem is s-RED/BLUE NONBLOCKER(s-R/B N): • Graph G=(V,E) of max degree s,V = VblueÈVred partition V • Is there a set V’ÍVred of size k s.t. every blue vertex has at least one neighbor not in V’?

  8. Lemma (cont’d) • Showing s-R/B NÎantimonotone W[1,s] • Easy • Õ(u blue)Sx_i a red neighbor of uØx_i • Showing W[1,s] <=fpt s-R/B N • Involved graph construction • Use blue vertices to limit weight of satisfying assignment

  9. Proof ofantimonotone W[1,s]=W[1,2] • Replaces literals with new variables representing subsets of literals of size 2..s • Replace inputs to OR gates with new variables • Adds additional variables and gates to enforce consistency (while keeping fan-in of OR gates bounded by 2) • Parameter k becomes k2k + Si=2..s C(k,i)

  10. Reducing WCNF-3SAT • So W[1,s]=antimonotone W[1,s]=W[1,2]=antimonotone W[1,2], • WCNF-3SATÎW[1,3],WCNF-2SAT- isantimonotone W[1,2]-hard • Paper’s result is much stronger, but: • WCNF-3SAT <=fpt WCNF-2SAT-

  11. Other results/implications • W[t]=antimonotone W[t] for t odd • Result that W[1]=Ès=1..∞ W[1,s] comes from LF(1,h) <=fpt W[1,s], s=f(h)(trade depth for fan-in) • IS, CLIQUE are W[1] – complete • Perfect Code, Weighted Exact CNF-SAT, Sized Subset Sum are W[1]-hard

  12. Conclusions • W[t] formulated uniformly for all t>=1 • W[1] stratification to Ès=1..∞ W[1,s] • W[1] collapses to W[1,2]

More Related