Homework 5: Due Nov 26, 2008

November 12th, 2008

PDF, and LaTeX tarball (with pics)

12 Responses to “Homework 5: Due Nov 26, 2008”

  1. Mike Clark Says:

    For 5.1.2, can we assume that the path the problem is talking about is a path with no cycles?

  2. Brian Matthews Says:

    For problem 4 what circuit gadgets can we use: AND, OR, XOR, NOT…?

  3. admin Says:

    Usually a path does not have cycles. If it does, and the cycles don’t overlap edges, this is called a walk.

    I’m not sure why you need circuit gadgets in problem 4: the goal is to reduce 3SAT to planar 3SAT. there are no circuit gadgets in either problem

  4. Brian Matthews Says:

    Then I guess if circuit pieces cannot be used to transmit the “signal”, then what are clauses that, I assume, connect verticies in the gadget?
    The reason I thought to use AND’s, OR’s was that they are mentioned in the diagram with the V between a1 and a_bar. By the way, what is the purpose of that diamond in figure 5.3?

  5. admin Says:

    well you have to use clauses as gadgets since you have to convert 3SAT to Planar 3SAT. The trick is finding the right clauses to use

  6. Brian Matthews Says:

    From what I saw in the book ANDs, ORs and NOTs are used, in figure 8.4 in the book on page 467. Seeing that a clause is only some inputs with some AND’s, OR’s, and NOT’s modifing on those inputs, I assume a set of variables (inputs) with some circles representing operations on some of those inputs is a gadget itself. It has only 1 output. The problem is constructing a circuit that is planar.

  7. John M Says:

    Is the graph G in an instance of LONGESTPATH a directed graph?

  8. admin Says:

    Not necessarily.

  9. John M Says:

    Are we allowed to use a directed version for the reduction part of the NPC proof?

  10. admin Says:

    You have to use the graph as specified. G (as stated) is undirected.

  11. Dafang Says:

    To generalize the reduction proof from an known NPC problem P1 to a target problem P2, is it correct to show that an arbitrary instance of P1 can be easily solved by solving P2 polynomial-order times rather than once?

  12. admin Says:

    In general yes

Leave a Reply