On-line Algorithm Homework2
Presenter: W. Choi
Problem: Exercise 10.4
Problem Description:
Prove the following 2-server algorithm ALG is O(1)-competitive in
any Euclidean Space.
ALG: After serving eac
Exercise 2.4 Consider the following generalization of RMTF. For any real
p [0,1] , let RMTFp be the algorithm that, upon a request for an item x,
moves x to the front with probability p. Generalize th
Exercise 3.3 Show that LRU does not incur Beladys anomaly but that
FIFO does incur the anomaly
Beladys Anomaly: Some reference strings generate more page faults when
more page frames are allotted.
1)
Homework 1
CSE 5314
Rui Huang
2/18/2004
Exercise 1.8 Let L be a list of two elements x and y. Prove that there is an optimal
offline algorithm OPT for L that satisfies the following properties: 1) OPT
Correctness of Work Function Algorithm for Metrical Task Systems
Usual dynamic programming table is constructed, but WFA selects one entry wi+1( si+1) in each row.
This is a state j that minimizes wi+
Exercise 1.7 (page 11)
For the static list accessing problem with l
items,
Instead of using a bound on the average
(over all initial configurations) static
optimal to derive 2l/(l+1) lower bound,
sh
Three Structurally Similar List Update Upper Bound Proofs
Review of amortized complexity:
ci = ci + (i) (i 1)
ci = ci + (i 1) + (i)
n
n
ci = ci + (0) ( n )
i=1
i=1
n
n
If (0) ( n ) is never positive,
CSE 5314 Lab Assignment 2
Due April 29, 2004
Goal:
Understanding of the work function algorithm for the k-server problem.
Requirements:
1.
Extend your dynamic programming solution to lab 1 to implemen
CSE 5314 Homework Set 1
Presentations Tentatively February 24 & 26
Reservations at http:/reptar.uta.edu/NOTES5314/hw1.txt
Each student should choose three problems from at least two different chapters
CSE 5314 Lab Assignment 1
Due March 9, 2004
Goals:
1.
Understanding of the k-server problem.
2.
Understanding of three approaches to the offline solution of the k-server problem: assignment problem (m