APPROX and RANDOM 2006 Online Algorithms to Minimize Resource Reallocation and Network Communication Sashka Davis, UCSD Jeff Edmonds, York University,

Slides:



Advertisements
Similar presentations
1+eps-Approximate Sparse Recovery Eric Price MIT David Woodruff IBM Almaden.
Advertisements

On allocations that maximize fairness Uriel Feige Microsoft Research and Weizmann Institute.
Coordination Mechanisms for Unrelated Machine Scheduling Yossi Azar joint work with Kamal Jain Vahab Mirrokni.
Minimum Clique Partition Problem with Constrained Weight for Interval Graphs Jianping Li Department of Mathematics Yunnan University Jointed by M.X. Chen.
 Review: The Greedy Method
Class-constrained Packing Problems with Application to Storage Management in Multimedia Systems Tami Tamir Department of Computer Science The Technion.
Online Algorithms Amrinder Arora Permalink:
Online Scheduling with Known Arrival Times Nicholas G Hall (Ohio State University) Marc E Posner (Ohio State University) Chris N Potts (University of Southampton)
Parallel Scheduling of Complex DAGs under Uncertainty Grzegorz Malewicz.
A polylogarithmic approximation of the minimum bisection Robert Krauthgamer The Hebrew University Joint work with Uri Feige.
1 Stochastic Event Capture Using Mobile Sensors Subject to a Quality Metric Nabhendra Bisnik, Alhussein A. Abouzeid, and Volkan Isler Rensselaer Polytechnic.
Item Pricing for Revenue Maximization in Combinatorial Auctions Maria-Florina Balcan, Carnegie Mellon University Joint with Avrim Blum and Yishay Mansour.
1 Approximation Algorithms for Demand- Robust and Stochastic Min-Cut Problems Vineet Goyal Carnegie Mellon University Based on, [Golovin, G, Ravi] (STACS’06)
Network Design Adam Meyerson Carnegie-Mellon University.
Institute of Computer Science University of Wroclaw Page Migration in Dynamic Networks Marcin Bieńkowski Joint work with: Jarek Byrka (Centrum voor Wiskunde.
Online Algorithms Motivation and Definitions Paging Problem Competitive Analysis Online Load Balancing.
An Approximation Algorithm for Requirement cut on graphs Viswanath Nagarajan Joint work with R. Ravi.
Online Oblivious Routing Nikhil Bansal, Avrim Blum, Shuchi Chawla & Adam Meyerson Carnegie Mellon University 6/7/2003.
Adaptiveness vs. obliviousness and randomization vs. determinism Dariusz Kowalski University of Connecticut & Warsaw University Andrzej Pelc University.
Ecole Polytechnique, Nov 11, List Scheduling on Related Machines processors Related machines: machines may have different speeds  0.25 
Algorithms and Economics of Networks: Coordination Mechanisms Abraham Flaxman and Vahab Mirrokni, Microsoft Research.
A General Approach to Online Network Optimization Problems Seffi Naor Computer Science Dept. Technion Haifa, Israel Joint work: Noga Alon, Yossi Azar,
Priority Models Sashka Davis University of California, San Diego June 1, 2003.
Price of Anarchy Bounds Price of Anarchy Convergence Based on Slides by Amir Epstein and by Svetlana Olonetsky Modified/Corrupted by Michal Feldman and.
A Model for Minimizing Active Processor Time Jessica Chang Joint work with Hal Gabow and Samir Khuller.
Minimizing Flow Time on Multiple Machines Nikhil Bansal IBM Research, T.J. Watson.
Online Function Tracking with Generalized Penalties Marcin Bieńkowski Institute of Computer Science, University of Wrocław, Poland Stefan Schmid Deutsche.
Scheduling Master - Slave Multiprocessor Systems Professor: Dr. G S Young Speaker:Darvesh Singh.
Minimizing Makespan and Preemption Costs on a System of Uniform Machines Hadas Shachnai Bell Labs and The Technion IIT Tami Tamir Univ. of Washington Gerhard.
1/24 Algorithms for Generalized Caching Nikhil Bansal IBM Research Niv Buchbinder Open Univ. Israel Seffi Naor Technion.
Approximation Algorithms for Stochastic Combinatorial Optimization Part I: Multistage problems Anupam Gupta Carnegie Mellon University.
International Graduate School of Dynamic Intelligent Systems, University of Paderborn Improved Algorithms for Dynamic Page Migration Marcin Bieńkowski.
Online Packet Admission and Oblivious Routing in Sensor Networks Mohamed Aly Department of Computer Science University of Pittsburgh And John Augustine.
Robust Network Supercomputing with Malicious Processes (Reliably Executing Tasks Upon Estimating the Number of Malicious Processes) Kishori M. Konwar*
Yossi Azar Tel Aviv University Joint work with Ilan Cohen Serving in the Dark 1.
THE STABILITY BOX IN INTERVAL DATA FOR MINIMIZING THE SUM OF WEIGHTED COMPLETION TIMES Yuri N. Sotskov Natalja G. Egorova United Institute of Informatics.
1 By: MOSES CHARIKAR, CHANDRA CHEKURI, TOMAS FEDER, AND RAJEEV MOTWANI Presented By: Sarah Hegab.
1 Introduction to Approximation Algorithms. 2 NP-completeness Do your best then.
Online Paging Algorithm By: Puneet C. Jain Bhaskar C. Chawda Yashu Gupta Supervisor: Dr. Naveen Garg, Dr. Kavitha Telikepalli.
1 Deterministic Auctions and (In)Competitiveness Proof sketch: Show that for any 1  m  n there exists a bid vector b such that Theorem: Let A f be any.
Department of Computer Science Stanley P. Y. Fung Online Preemptive Scheduling with Immediate Decision or Notification and Penalties.
Chapter 8 PD-Method and Local Ratio (4) Local ratio This ppt is editored from a ppt of Reuven Bar-Yehuda. Reuven Bar-Yehuda.
Lower Bounds in Greedy Model Sashka Davis Advised by Russell Impagliazzo (Slides modified by Jeff) UC San Diego October 6, 2006.
Facility Location in Dynamic Geometric Data Streams Christiane Lammersen Christian Sohler.
A polylog competitive algorithm for the k-server problem Nikhil Bansal (IBM) Niv Buchbinder (Open Univ.) Aleksander Madry (MIT) Seffi Naor (Technion)
Hierarchical Well-Separated Trees (HST) Edges’ distances are uniform across a level of the tree Stretch  = factor by which distances decrease from root.
A Optimal On-line Algorithm for k Servers on Trees Author : Marek Chrobak Lawrence L. Larmore 報告人:羅正偉.
Frequency Capping in Online Advertising Moran Feldman Technion Joint work with: Niv Buchbinder,The Open University of Israel Arpita Ghosh,Yahoo! Research.
1 Chapter 5-1 Greedy Algorithms Slides by Kevin Wayne. Copyright © 2005 Pearson-Addison Wesley. All rights reserved.
© The McGraw-Hill Companies, Inc., Chapter 12 On-Line Algorithms.
Non-Preemptive Buffer Management for Latency Sensitive Packets Moran Feldman Technion Seffi Naor Technion.
International Graduate School of Dynamic Intelligent Systems, University of Paderborn Fighting Against Two Adversaries: Page Migration in Dynamic Networks.
Approximating Buy-at-Bulk and Shallow-Light k-Steiner Trees Mohammad T. Hajiaghayi (CMU) Guy Kortsarz (Rutgers) Mohammad R. Salavatipour (U. Alberta) Presented.
1 Approximation Algorithms for Generalized Scheduling Problems Ravishankar Krishnaswamy Carnegie Mellon University joint work with Nikhil Bansal, Anupam.
Eric Allender Rutgers University Curiouser and Curiouser: The Link between Incompressibility and Complexity CiE Special Session, June 19, 2012.
11 -1 Chapter 12 On-Line Algorithms On-Line Algorithms On-line algorithms are used to solve on-line problems. The disk scheduling problem The requests.
NOTE: To change the image on this slide, select the picture and delete it. Then click the Pictures icon in the placeholder to insert your own image. Fast.
Common Intersection of Half-Planes in R 2 2 PROBLEM (Common Intersection of half- planes in R 2 ) Given n half-planes H 1, H 2,..., H n in R 2 compute.
Models of Greedy Algorithms for Graph Problems Sashka Davis, UCSD Russell Impagliazzo, UCSD SIAM SODA 2004.
The Price of Routing Unsplittable Flow Yossi Azar Joint work with B. Awerbuch and A. Epstein.
CSE 5314 – Spring 2004 Homework #1 - Solutions [Soumya Sanyal] Q 1.5: Prove that the algorithm “MTF every-other-access” (i.e. move the item to the front.
Errol Lloyd Design and Analysis of Algorithms Approximation Algorithms for NP-complete Problems Bin Packing Networks.
Chapter 8 PD-Method and Local Ratio (5) Equivalence This ppt is editored from a ppt of Reuven Bar-Yehuda. Reuven Bar-Yehuda.
Scheduling Determines the precise start time of each task.
Multiprocessor Real-Time Scheduling
Instructor: Shengyu Zhang
A new and improved algorithm for online bin packing
University of Pittsburgh
György Dósa – M. Grazia Speranza – Zsolt Tuza:
Clustering.
Presentation transcript:

APPROX and RANDOM 2006 Online Algorithms to Minimize Resource Reallocation and Network Communication Sashka Davis, UCSD Jeff Edmonds, York University, Canada Russell Impagliazzo, UCSD

APPROX and RANDOM 2006 Resource Allocation Problems [KKD02, PL95, IRSD99, Edm00] Given: Multi-processor machine with T identical processors. Problem: assign processors to parallel jobs whose requirements are evolving and malleable. Goal: schedule jobs, satisfy processor requirements of each job, minimize preemption.

APPROX and RANDOM 2006 The Weak Department Chair Problem I want 12!

APPROX and RANDOM 2006 RAP: Resource Allocation Problem RAP Instance T identical processors. n users. Input: (i,r t,i ) - at time t user i requests r i,t processors. Output: (l t,i ) - the algorithm must allocate l t,i processors to i, l t,i ≥ r t,i. Constraints: ∑ r t,i ≤ T and ∑ l t,i ≤ T, for all t. Objective: Minimize changes to the global state. Cost = |{(l t,i,l t+1,i ), where l t,i ≠ l t+1,i }|. The algorithm is not notified when users current demands fall bellow their current allocations.

APPROX and RANDOM 2006 The Strong Department Chair Problem I want 30, If not – penalty! You can’t have 30! I take the penalty!

APPROX and RANDOM 2006 RAPP: Resource Allocation Problem with Penalties RAPP Instance T identical processors. n users. Input: (i,r t,i, p t,i ) - at time t user i requests r t,i processors and penalty p t,i. Output: (l t,i ) - allocation of l t,i, processors to i s.t., l t,i ≥ r t,i or do nothing. Constraints: ∑ r t,i ≤ T and ∑ l t,i ≤ T, for all t. Objective: Minimize changes to the global state, i.e., reallocations. Cost: |{(l t,i,l t+1,i ), where l t,i ≠ l t+1,i }| + ∑ p t,i, when the scheduler fails to satisfy the t’th request. The algorithm is not notified when its current demand falls bellow its current allocation.

APPROX and RANDOM 2006 The Humble Chair Problem I want MORE ! ?

APPROX and RANDOM 2006 RRAP: Restricted Resource Allocation Problem RRAP Instance T identical processors n users Input: (i) - at time t user i complains. Output: (l t,i ), such that l t,i ≥ l t-1,i. Constraints: ∑ l j,t ≤ T, for all t. Objective: Minimize changes to the global state, i.e., reallocations. Cost: |{(l t,i,l t+1,i )}|, such that l t,i ≠ l t+1,i. The algorithm never learns the precise demands exactly, only an upper bound for each. ?

APPROX and RANDOM 2006 Network Communication Problem [OLW01, CKA02, CYV06 ] Central cache and a network of low-power sensors. Sensors read values. 1.Cache must know the values read exactly – #sensor reads = #network transmissions. 2.Sensors are low-power devices and we want to minimize network communication. –Solution: Settle for approximation.

APPROX and RANDOM 2006 TMAV: Transmission Minimizing Approximate Value Problem n sensors reading values Sensor 1 [L 1,,H 1 ] Sensor 1 [L 1,,H 1 ] Sensor n [L n,H n ] Sensor n [L n,H n ] v n  [L n,H n ] v1v1 v 1  [L′ 1, H′ 1, ] Precision T ≥ ∑(H i -L i ) Constraints: T ≥ ∑(H i -L i ); v i  [L i,H i ], for all t, i Objective: Minimize network communication. Cost: The number of transmissions between sensors and cache. Central Cache

APPROX and RANDOM 2006 Two Online Problems Minimize Resource Reallocation Minimize Network Communication Central Control Maintains State. Must satisfy the demands of many users. Objective: Minimize changes to the state. A property: online algorithms do NOT know the precise requirements of users. TMAV ? RRAP RAPPRAP

APPROX and RANDOM 2006 Bi-criteria Online Algorithms Adversary uses T resources/precision. Algorithm: –use sT resources/precision. –the precise requirements of users are unknown to the algorithm. Goal: Find randomized, competitive online algorithms for RAP, RRAP, RAPP, and TMAV problems using the smallest possible s. When s=1 then the competitive ratio is infinity.

APPROX and RANDOM 2006 Results: Upper Bounds 1.O(log s n)-competitive algorithm for RRAP, where s is a constant, s≥3. 2.Modified the solution for RRAP and obtained algorithms with similar competitive ratios O(log s n) for RAP, RAPP, and TMAV. ?

APPROX and RANDOM 2006 Results: Lower Bounds 1.For s = 1 no competitive algorithm for RAP and TMAV exists. 2.Defined the notion of competitive ratio preserving online reduction with respect to adaptive online adversary “≤ AD_ON ’’. 1.RAP ≤ AD_ON TMAV 1.RAP ≤ AD_ON RAPP

APPROX and RANDOM 2006 Results: Lower Bounds Using Reductions (h,k)-paging ≤ AD_ON RAP 1.No online algorithm, using (1+ε) resources can achieve competitive ratio better than Ω(1/ ε) against an adaptive online adversary, using resource of size 1. 2.No online algorithm using (1+ ε) resources can achieve competitive ratio better than Ω(log(1/ ε)) against an oblivious adversary using resource of size 1.

APPROX and RANDOM 2006 The Remainder of the Talk 1.Steal From the Rich – a randomized O(log s n)-competitive algorithm for RRAP. 2.For s=1 no competitive algorithm for RAP and TMAV exists.

APPROX and RANDOM 2006 RRAP: Restricted Resource Allocation Problem RRAP Instance : T identical processors, n users. Input: (i) - at time t user i complains. Output: (l i,t ), such that l t,i ≥ l t-1,i. Constraints: ∑ l t,i ≤ T, for all t. Cost: Number of pairs (l t,i,l t+1,i ), such that l t,i ≠ l t+1,i. The algorithm never learns the precise demands exactly, only an upper bound for each. ?

APPROX and RANDOM 2006 Steal From the Rich Algorithm sT/n user 1 sT/n user 2 sT/n user n Initially partition sT resources evenly among the n users. Let s be a constant, and r=Θ(√s), μ be a constants, which depend on s, but not the instance.

APPROX and RANDOM 2006 Steal From the Rich Algorithm l t,1 user 1 l t,2 user 2 l t,k user k At time t+1 user j complains. l t,j user j SFR picks a user k from [n]-{j} with probability l t,k /(sT-l t,j ). l t,n user n δ l t+1,k ← l t,k -δ; l t+1,,j +1←l t,j +δ; δ l t,j user j user k μT/n SFROPT

APPROX and RANDOM 2006 How Much to Steal from the Rich? SFR maintains the following invariants: 1.All users have at least μT/n l t+1,k ≥ μT/n, hence δ ≤ l t,k - μT/n; 2.l t+1,k does not shrink by a factor more than 1/r l t+1,k ≥ l k,t /r, hence δ ≤ l k,t (r-1)/r; 3.l t+1,j does not grow by a factor more than r l t+1,j ≤ rl t,j,, hence δ ≤ l j,t (r-1); δ = min {l t,k -μT/n; l t,k (r-1)/r; l t,j (r-1)}.

APPROX and RANDOM 2006 SFR Analysis Want to show that for any req. sequence σ E(SFRs(σ)) ≤ O(log s n)OPT(σ)+d. Φ: R n  R n → R + ; a t =SFR t +(Φ t -Φ t-1 ) E(SFRs(σ)) = E(∑SFR t )=E(∑a t )-Φ end +Φ 0 Want to prove that for all t: Φ t ≤ O(n log s n), for all t, E(a t ) ≤ O(log s n)OPT t. Then Φ 0 ≤ O(n log s n), and we use d = O(n log s n).

APPROX and RANDOM 2006 SFR Potential Function ΔΦ is small when SFR and OPT have proportional allocations. When SFR has cost and OPT does not, then ΔΦ is negative and compensates for the actual cost of SFR.

APPROX and RANDOM 2006 Amortized Update Cost E(a t ) = E(SFR t + ΔΦ t ) ≤ O(log s n)OPT t Case 1: OPT t ≠ 0, SFR = 0. E(a t ) = E(0 + #changed intervals  O(log s n)) ≤ O(log s n)OPT t Case 2: OPT t = 0, SFR = 2. E(a t ) = E(2+ΔΦ t ) E(ΔΦ t ) ≤ -2. In Case 2, SFR does: –l t,j grows by a factor of r then ΔΦ t )≤-14; –l t,k shrinks by a factor of 1/r then ΔΦ t ≤-14; –Neither: (δ = l t,k -μT/n) then ΔΦ t ≥ 0 (unfortunate but rare event). Concluding: E(SFRs(σ)) ≤ O(log s n)OPT(σ)+d.

APPROX and RANDOM 2006 The Additional Resource is Vital Theorem: There is no online algorithm using T resources that is f(n) competitive against and adversary using T resources, for any function f. Consider RAP with 2 users and T=1.

APPROX and RANDOM 2006 If s=1 then competitive ratio is ∞ 0 1 user1user2 1.Adversary cost is 2. 2.Probability of incurring cost during t’th request is 1/8t. 3.The expected cost of the algorithm diverges as t goes to infinity. S 4,1 <rS 2,1 <r S 1,1 < r r  [0,1] S 4,1 <rS≥r S 3,2 = 1-S

APPROX and RANDOM 2006 Relating the Hardness of the Problems TMAV RAP RAPP SFR RRAP ≤ AD_ON ? SFR

APPROX and RANDOM 2006 Conclusions 1.We obtained O(log s n)-competitive algorithms for four different problems. 2.Justified the need for sT resource. Defined a notion of online reduction with respect to adaptive online adversary. Related the hardness of the problems using online reductions. Reduced (h-k)-Paging to RAP and transferred the standard paging lower bounds to the four problems.

APPROX and RANDOM 2006 New Issues We studied memoryless online algorithms that do not know the current demands exactly. Online reductions to leverage existing lower bounds and relate hardness of online problems.

APPROX and RANDOM 2006 Open problems Close the gap between the upper and lower bounds. Can competitive ratio preserving reductions with respect to adaptive online adversary deliver other lower bounds for other problems? Do other problems have similar memoryless online solutions, where the algorithm does not know the demands exactly, but only an upper bound approximation of it.