2ima00

This shows you the differences between two versions of the page.

Both sides previous revision Previous revision Next revision | Previous revision Next revision Both sides next revision | ||

2ima00 [2016/05/26 11:42] bmpjansen [Student lectures] |
2ima00 [2016/05/26 12:18] bmpjansen [Student lectures] |
||
---|---|---|---|

Line 124: | Line 124: | ||

**Kernels: Chris & Leo.** | **Kernels: Chris & Leo.** | ||

+ | |||

+ | The following paper gives a kernel that is not as small, but is easier to compute, than the one given in the book. In particular, read Section 4 of this paper: | ||

[[http://dx.doi.org/10.1007/s00224-009-9234-2| | [[http://dx.doi.org/10.1007/s00224-009-9234-2| | ||

Line 129: | Line 131: | ||

A Cubic Kernel for Feedback Vertex Set and Loop Cutset. Theory Comput. Syst. 46(3): 566-597 (2010)]] | A Cubic Kernel for Feedback Vertex Set and Loop Cutset. Theory Comput. Syst. 46(3): 566-597 (2010)]] | ||

- | This uses an approximation algorithm, which was developed independently by 2 sets of authors. Pick the one you find the easiest to read: | + | Instead of requiring an algorithm to find maximum matchings, it just needs an approximation algorithm for (weighted) feedback vertex set. It was developed independently by 2 sets of authors. Pick the one you find the easiest to read: |

Option 1: | Option 1: | ||

Line 141: | Line 143: | ||

**Treewidth: Henk & Xi.** | **Treewidth: Henk & Xi.** | ||

+ | |||

+ | There are 2 approaches for solving the problem when you have a tree decomposition. The first one has a running time of 3^k * poly(n), harder to understand, but may be easier to implement: | ||

[[https://arxiv.org/pdf/1103.0534v1.pdf|Cygan et al.: Solving connectivity problems parameterized by treewidth in single exponential time]] | [[https://arxiv.org/pdf/1103.0534v1.pdf|Cygan et al.: Solving connectivity problems parameterized by treewidth in single exponential time]] | ||

+ | |||

+ | The next approach has a worse factor f(k), but a better polynomial term (linear), and is conceptually simpler: | ||

+ | |||

[[http://link.springer.com/chapter/10.1007%2F3-540-36379-3_25|Ton Kloks, C.M. Lee, Jiping Liu. New Algorithms for k-Face Cover, k-Feedback Vertex Set, and k-Disjoint Cycles on Plane and Planar Graphs]] (page 8 and further) | [[http://link.springer.com/chapter/10.1007%2F3-540-36379-3_25|Ton Kloks, C.M. Lee, Jiping Liu. New Algorithms for k-Face Cover, k-Feedback Vertex Set, and k-Disjoint Cycles on Plane and Planar Graphs]] (page 8 and further) | ||

+ | |||

+ | I don't know which of the 2 approaches will be faster on the inputs we have; one has a worse polynomial term, the other has a worse factor f(k). I don't think it is feasible to implement both, given the time. Pick one of the two and go for it. | ||

**Parameterized by solution size (iterative compression & randomized): Huib & Stefan.** | **Parameterized by solution size (iterative compression & randomized): Huib & Stefan.** | ||

+ | |||

+ | For the iterative compression algorithm, I think the book gives sufficient details to build an implementation from. There is a theoretically faster algorithm, which is linked below. You could read it for inspiration, but it's not necessary and I do not think it is feasible to implement it fully because it needs a subroutine with matroid computations. | ||

[[http://dx.doi.org/10.1016/j.ipl.2014.05.001|Tomasz Kociumaka, Marcin Pilipczuk: | [[http://dx.doi.org/10.1016/j.ipl.2014.05.001|Tomasz Kociumaka, Marcin Pilipczuk: | ||

Faster deterministic Feedback Vertex Set. Inf. Process. Lett. 114(10): 556-560 (2014)]] | Faster deterministic Feedback Vertex Set. Inf. Process. Lett. 114(10): 556-560 (2014)]] | ||

+ | |||

+ | For the randomized algorithm, I don't know of any additional literature. You could search by yourself. |

2ima00.txt ยท Last modified: 2016/05/26 12:27 by bmpjansen