CSC 7103 20091008

CSC 7103 20091008 - f are all overloaded. We dont know...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon
CSC 7103 10/8/2009 Page 1 of 2 Type up notes from notepad. Claim : SF A (σ) has a competitive ratio of 2 and doesn’t fail. M is # of machines, N is # of jobs Proof : We must show that if you keep following the algorithm, then by the time you got to the last job, you would be able to assign that job to a machine. Consider jobs r 1 , r 2 , …, r N-1 , r N At the last job we must show there is no machine that fits the criteria for the algorithm (in choosing the machine). We divide the set of machines into two sets, one of lightly loaded machines and one of heavily loaded machines. Figure 1 – machine split Let f be the index of the fastest machine whose load l N-1 (f) <= OPT(σ) <= A. What can we say about l N-1 (M)? We must show that if f exists, f must be < M and all nodes after M
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Background image of page 2
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: f are all overloaded. We dont know anything about the load of nodes before M f . Figure 2 machine knowledge B = { i | i &gt; f } C = { all other machines not in B } S i = Set of jobs assigned to machine by slow fit to machine i in the set B S i * = Set of jobs assigned to machine by OPT() to machine i in the set B If S i * = S i , then OPT behaved the same as slow fit, so we need to show there is a difference in the sets. Suppose B = { 1, 2, , M , then all machines are overloaded and this is undesirable. To show that f exists we must show that B is not the set of all machines. Proof of slow fit algorithm CSC 7103 10/8/2009 Page 2 of 2 Static scheduling...
View Full Document

Page1 / 2

CSC 7103 20091008 - f are all overloaded. We dont know...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online