Coupled Bisection for Root-Ordering

Size: px
Start display at page:

Download "Coupled Bisection for Root-Ordering"

Transcription

1 Peter I. Frazier Shane G. Henderson Cornell University ORIE Cornell University ORIE Cornell University ORIE July 17, 2015

2 One-Dimensional Root-Finding 0 1 Problem Statement Structural Results Bisection Policy 2/28

3 One-Dimensional Root-Finding f 0 1 x Problem Statement Structural Results Bisection Policy 2/28

4 One-Dimensional Root-Finding f 0 1 x x Problem Statement Structural Results Bisection Policy 2/28

5 Coupled Root-Ordering 0 1 Problem Statement Structural Results Bisection Policy 3/28

6 Coupled Root-Ordering f(s 1, ) f(s 2, ) f(s 3, ) 0 1 x (s 1 ) x (s 2 ) x (s 3 ) Problem Statement Structural Results Bisection Policy 3/28

7 Coupled Root-Ordering f(s 1, ) f(s 2, ) f(s 3, ) 0 1 x (s 1 ) x (s 2 ) x (s 3 ) x Problem Statement Structural Results Bisection Policy 3/28

8 Coupled Root-Ordering Let S denote a set of elements (large but finite). Let f : S [0, 1) R, monotonic in its second argument. For every element s S, denote x (s) as the unique root of f (s, ). Suppose when we evaluate f for a given x [0, 1), we find f (s, x) for every s S. Goal: find the complete ordering of elements s by their roots x (s) with the least number of function evaluations as possible. Problem Statement Structural Results Bisection Policy 4/28

9 Applications Multi-armed Bandit Problems Sorting states in state space according to their Gittens or Whittle indices. Ordering of states is all that is required to implement optimal policy. Remote Sensing E.g., finding the boundary of a forest from an airborne laser scanner. Problem Statement Structural Results Bisection Policy 5/28

10 The Naive Approach Problem Statement Structural Results Bisection Policy 6/28

11 The Naive Approach Problem Statement Structural Results Bisection Policy 6/28

12 The Naive Approach Problems with Discretizing the Interval Non-adaptive in nature; unclear how to determine subinterval length. May require more computational work than necessary. Problem Statement Structural Results Bisection Policy 6/28

13 The Clever Approach Problem Statement Structural Results Bisection Policy 7/28

14 The Clever Approach 4 Problem Statement Structural Results Bisection Policy 7/28

15 The Clever Approach 2 2 Problem Statement Structural Results Bisection Policy 7/28

16 The Clever Approach Problem Statement Structural Results Bisection Policy 7/28

17 The Clever Approach How can we adaptively choose where to query f? Problem Statement Structural Results Bisection Policy 7/28

18 Formal Definition State Variables 2 2 Problem Statement Structural Results Bisection Policy 8/28

19 Formal Definition State Variables n (0) n (1) 2 2 x (0) x (1) x (2) Problem Statement Structural Results Bisection Policy 8/28

20 Formal Definition State Variables n (0) n (1) n (2) x (0) x (1) x (2) x (3) Problem Statement Structural Results Bisection Policy 8/28

21 Formal Definition State Variables n (0) n (1) n (2) x (0) x (1) x (2) x (3) At time epoch j: { } Let X j = [x (0), x (1) ),..., [x (j), x (j+1) ) denote the current partition of the interval. Let N j = (n (0), n (1),..., n (j) ) denote the number of roots in each subinterval. The pair (X j, N j ) denotes the computational state. Problem Statement Structural Results Bisection Policy 8/28

22 Formal Definition Transition We assume i.i.d. priors on x (s) for every element s S. Without loss of generality, x (s) Uniform[0, 1). Problem Statement Structural Results Bisection Policy 9/28

23 Formal Definition Transition We assume i.i.d. priors on x (s) for every element s S. Without loss of generality, x (s) Uniform[0, 1). Problem Statement Structural Results Bisection Policy 9/28

24 Formal Definition Transition We assume i.i.d. priors on x (s) for every element s S. Without loss of generality, x (s) Uniform[0, 1). Problem Statement Structural Results Bisection Policy 9/28

25 Formal Definition Transition We assume i.i.d. priors on x (s) for every element s S. Without loss of generality, x (s) Uniform[0, 1). 2 2 Problem Statement Structural Results Bisection Policy 9/28

26 Formal Definition Transition We assume i.i.d. priors on x (s) for every element s S. Without loss of generality, x (s) Uniform[0, 1). 2 2 N j+1 = ( n (0),..., n (l 1), N (l), n (l) N (l), n (l+1),..., n (j)), where N (l) Binomial ( n (l), x x (l) ) x (l+1) x (l). Problem Statement Structural Results Bisection Policy 9/28

27 Formal Definition Objective Function Problem Statement Structural Results Bisection Policy 10/28

28 Formal Definition Objective Function Problem Statement Structural Results Bisection Policy 10/28

29 Formal Definition Objective Function Problem Statement Structural Results Bisection Policy 10/28

30 Formal Definition Objective Function Let τ denote the number of function evaluations to sort all elements, i.e., τ = inf{j N : n (i) j 1 i}. Problem Statement Structural Results Bisection Policy 10/28

31 Formal Definition Defining the Optimal Policy Problem Statement Structural Results Bisection Policy 11/28

32 Formal Definition Defining the Optimal Policy A policy π is a mapping from computational states (X, N ) to a refined partition X that adds one additional subinterval. Problem Statement Structural Results Bisection Policy 11/28

33 Formal Definition Defining the Optimal Policy A policy π is a mapping from computational states (X, N ) to a refined partition X that adds one additional subinterval. Let W π (X, N ) denote the expected number of function evaluations required to sort elements, where we start at state (X, N ), i.e., W π (X, N ) = E π [τ (X 0, N 0 ) = (X, N )]. Problem Statement Structural Results Bisection Policy 11/28

34 Formal Definition Defining the Optimal Policy A policy π is a mapping from computational states (X, N ) to a refined partition X that adds one additional subinterval. Let W π (X, N ) denote the expected number of function evaluations required to sort elements, where we start at state (X, N ), i.e., W π (X, N ) = E π [τ (X 0, N 0 ) = (X, N )]. Let W (X, N ) = inf π W π (X, N ) denote the expected number of evaluations required under the optimal policy. Problem Statement Structural Results Bisection Policy 11/28

35 Dynamic Programming By the principle of optimality, we can iteratively take the best partition X, giving us a recursion. Problem Statement Structural Results Bisection Policy 12/28

36 Dynamic Programming By the principle of optimality, we can iteratively take the best partition X, giving us a recursion. Theorem (Basic Recursion) For a given computational state (X, N ), we have W (X, N ) = 1 + inf X [ E W ( X, N ] ) (X 0, N 0 ) = (X, N ), with stopping conditions W (X, N ) = 0 if n (i) 1 for all i. Problem Statement Structural Results Bisection Policy 12/28

37 Divide and Conqueor Theorem (Decomposition and Interval Invariance) Under the optimal policy, the value function W has the following two properties. Decomposition: W (X, N ) = N 1 i=0 W Interval Invariance: ({ } W [x (i), x (i+1) ), n (i)) = W ({ } [x (i), x (i+1) ), n (i)) ( {[0, 1)}, n (i)). Problem Statement Structural Results Bisection Policy 13/28

38 Decomposition Ex. Expected remaining number of iterations required to finish sorting 7 elements: ({ } ) W [x (0), x (1) ), [x (1), x (2) ), [x (2), x (3) ), (2, 2, 3) Problem Statement Structural Results Bisection Policy 14/28

39 Decomposition Ex. Expected remaining number of iterations required to finish sorting 7 elements: ({ } ) W [x (0), x (1) ), [x (1), x (2) ), [x (2), x (3) ), (2, 2, 3) }{{} sort 2 elements }{{} sort 2 elements } {{ } sort 3 elements Problem Statement Structural Results Bisection Policy 14/28

40 Decomposition Ex. Expected remaining number of iterations required to finish sorting 7 elements: ({ } ) W [x (0), x (1) ), [x (1), x (2) ), [x (2), x (3) ), (2, 2, 3) }{{} sort 2 elements }{{} sort 2 elements } {{ } sort 3 elements = W ( {[x (0), x (1) )}, 2 ) + W ( {[x (1), x (2) )}, 2 ) + W ( {[x (2), x (3) )}, 3 ) Problem Statement Structural Results Bisection Policy 14/28

41 Interval Invariance 3 a b Problem Statement Structural Results Bisection Policy 15/28

42 Interval Invariance 3 a 1 2 b a b Problem Statement Structural Results Bisection Policy 15/28

43 Interval Invariance 3 a 1 2 b a b a b Problem Statement Structural Results Bisection Policy 15/28

44 Interval Invariance 3 a 1 2 b a b a b a b Problem Statement Structural Results Bisection Policy 15/28

45 Interval Invariance 3 a 1 2 b a b Re-scaling a b a b Problem Statement Structural Results Bisection Policy 15/28

46 Interval Invariance 3 3 a 1 2 b a b Re-scaling a b a b 0 1 Problem Statement Structural Results Bisection Policy 15/28

47 A Simpler Recursion n 0 1 Problem Statement Structural Results Bisection Policy 16/28

48 A Simpler Recursion n 0 1 N x Binomial(n, x) N x n N x 0 x x 1 Problem Statement Structural Results Bisection Policy 16/28

49 A Simpler Recursion n 0 W ({[0, 1)}, n) 1 N x Binomial(n, x) N x n N x 0 x x 1 W ({[0, x)}, N x ) + W ({[x, 1)}, n N x ) Problem Statement Structural Results Bisection Policy 16/28

50 A Simpler Recursion n 0 W ({[0, 1)}, n) 1 N x Binomial(n, x) N x n N x W ({[0, 1)}, N x ) + W ({[0, 1)}, n N x ) Problem Statement Structural Results Bisection Policy 16/28

51 A Simpler Recursion n 0 W (n) 1 N x Binomial(n, x) N x n N x W (N x ) + W (n N x ) Problem Statement Structural Results Bisection Policy 16/28

52 A Simpler Recursion For notation purposes, let W (n) := W ({[0, 1)}, n). Problem Statement Structural Results Bisection Policy 16/28

53 A Simpler Recursion For notation purposes, let W (n) := W ({[0, 1)}, n). Corollary (Simple Recursion) W (n) = 1 + min x (0,1) E [W (N x) + W (n N x )], where N x Binomial (n, x), and W (0) = W (1) = 0. Problem Statement Structural Results Bisection Policy 16/28

54 Computing the Optimal Policy W (n) = 1 + min x (0,1) E [W (N x) + W (n N x )] Problem Statement Structural Results Bisection Policy 17/28

55 Computing the Optimal Policy W (n) = 1 + min x (0,1) E [W (N x) + W (n N x )] Using the stopping conditions W (0) = W (1) = 0, we can iteratively compute W ( ) for arbitrary n. Problem Statement Structural Results Bisection Policy 17/28

56 Computing the Optimal Policy W (n) = 1 + min x (0,1) E [W (N x) + W (n N x )] Using the stopping conditions W (0) = W (1) = 0, we can iteratively compute W ( ) for arbitrary n. Problem: N x {0, n} with positive probability. Problem Statement Structural Results Bisection Policy 17/28

57 Computing the Optimal Policy W (n) = 1 + min x (0,1) E [W (N x) + W (n N x )] Using the stopping conditions W (0) = W (1) = 0, we can iteratively compute W ( ) for arbitrary n. Problem: N x {0, n} with positive probability. This is an implicit definition of W. Problem Statement Structural Results Bisection Policy 17/28

58 Computing the Optimal Policy Theorem (Computational Recursion) For all n 2, we have { 1 W (n) = min x (0,1) 1 x n (1 x) n [ ]} + E W (N x ) + W (n N x ) 1 N x n 1, where N x Binomial (n, x), and W (0) = W (1) = 0. Problem Statement Structural Results Bisection Policy 17/28

59 Bisection Policy Suppose we queried the midpoint of the interval, regardless of the number of elements to sort, i.e., choose x = 1/2 for all n. We call this the Bisection Policy. Denote W β (n) the expected number of function evaluations to sort n elements by their roots, under the bisection policy. Problem Statement Structural Results Bisection Policy 18/28

60 Bisection Policy Theorem (Bisection Policy Recursion) For all n 2, we have [ ] W β (n) = E W β (N 1/2 ), where N 1/2 Binomial (n, 1/2), and W β (0) = W β (1) = 0. Problem Statement Structural Results Bisection Policy 18/28

61 Bisection vs. Optimal Policy 3000 W β (n) (Bisection Policy) W(n) (Optimal Policy) 2500 Expected Function Evaluations n Problem Statement Structural Results Bisection Policy 19/28

62 Is Bisection Optimal? Problem Statement Structural Results Bisection Policy 20/28

63 Is Bisection Optimal? No. Consider the case n = 6. Problem Statement Structural Results Bisection Policy 20/28

64 Is Bisection Optimal? The case n = Expected Function Evaluations x Problem Statement Structural Results Bisection Policy 20/28

65 Is Bisection Optimal? Linear Growth Rate W β (n) (Bisection Policy) W(n) (Optimal Policy) Linear Rate n Problem Statement Structural Results Bisection Policy 20/28

66 Theorem (Lower Bound) For all n 0, we have W (n) n 1. Problem Statement Structural Results Bisection Policy 21/28

67 Theorem (Lower Bound) For all n 0, we have W (n) n 1. Proof : If we want to sort n elements, we must perform at least n 1 function evaluations of f to seperate them. Problem Statement Structural Results Bisection Policy 21/28

68 Theorem (Lower Bound) For all n 0, we have W (n) n 1. Proof : If we want to sort n elements, we must perform at least n 1 function evaluations of f to seperate them. Can we find upper bounds on W β? Problem Statement Structural Results Bisection Policy 21/28

69 What s the difference? For notation purposes, define h(n) = h(n + 1) h(n). Theorem (Recursion for W β ) For all n 2, we have [ ] W β (n) = E W β (N 1/2 ), where N 1/2 Binomial (n, 1/2), and stopping conditions W β (0) = 0 and W β (1) = 2. Problem Statement Structural Results Bisection Policy 22/28

70 Bounding the Bisection Policy Goal: show for some m, γ > 0 that W β (n) γ for all n m. Problem Statement Structural Results Bisection Policy 23/28

71 Bounding the Bisection Policy Goal: show for some m, γ > 0 that W β (n) γ for all n m. Each subsequent term of W β is a weighted average of the previous terms in the sequence. Idea: Induction! Problem Statement Structural Results Bisection Policy 23/28

72 Bounding the Bisection Policy Goal: show for some m, γ > 0 that W β (n) γ for all n m. E [ W β (N 1/2 ) ] E [ W β (N 1/2 ) N 1/2 m 1 ] E [ W β (N 1/2 ) N 1/2 m ] Bound above using? Bound above using Induction Hypthosis Problem Statement Structural Results Bisection Policy 23/28

73 Computational Upper Bounds Growth Rate of Expected Evaluations Required to Sort W β (n) (Linear Growth Rate) PMF of Truncated Binonmial Distribution Problem Statement Structural Results Bisection Policy 24/28

74 Computational Upper Bounds Growth Rate of Expected Evaluations Required to Sort W β (n) (Linear Growth Rate) PMF of Truncated Binonmial Distribution Problem Statement Structural Results Bisection Policy 24/28

75 Computational Upper Bounds Growth Rate of Expected Evaluations Required to Sort W β (n) (Linear Growth Rate) PMF of Truncated Binonmial Distribution Problem Statement Structural Results Bisection Policy 24/28

76 Computational Upper Bounds Growth Rate of Expected Evaluations Required to Sort W β (n) (Linear Growth Rate) PMF of Truncated Binonmial Distribution Problem Statement Structural Results Bisection Policy 24/28

77 A Useful Property of Binomial RVs Problem Statement Structural Results Bisection Policy 25/28

78 A Useful Property of Binomial RVs Problem Statement Structural Results Bisection Policy 25/28

79 A Useful Property of Binomial RVs Problem Statement Structural Results Bisection Policy 25/28

80 A Useful Property of Binomial RVs Problem Statement Structural Results Bisection Policy 25/28

81 Computational Upper Bounds Growth Rate of Expected Evaluations Required to Sort W β (n) (Linear Growth Rate) PMF of Truncated Binonmial Distribution Problem Statement Structural Results Bisection Policy 26/28

82 Computational Upper Bounds Growth Rate of Expected Evaluations Required to Sort W β (n) (Linear Growth Rate) Monotone Decreasing Upper Bound PMF of Truncated Binonmial Distribution Problem Statement Structural Results Bisection Policy 26/28

83 Computational Upper Bounds Theorem (Upper Bound on Bisection Policy) For some non-negative integer m, we define g m (l) = max k [l,m 1] W β (k), and define h as ( ) ] h m (n) = E [g m N 1/2 N1/2 m 1, where N 1/2 Binomial (n, 1/2). Suppose we have γ m > 0 so that the following condition holds: { } max W β (m), h m (m) γ m. Then for all n m, it must be that W β (n) γ m. Problem Statement Structural Results Bisection Policy 26/28

84 Computing Upper Bounds Use the computational recursion to compute W β (k) for k = {1, 2,..., m} for some positive integer m. Compute γ m = max{ W β (m), h m (m)}. Problem Statement Structural Results Bisection Policy 27/28

85 Computing Upper Bounds Use the computational recursion to compute W β (k) for k = {1, 2,..., m} for some positive integer m. Compute γ m = max{ W β (m), h m (m)}. Corollary (Bounds for Bisection Policy) For n 15, we have that n 1 W (n) W β (n) γ 15 (n 15) + W β (15), where γ 15 = and W β (15) = Problem Statement Structural Results Bisection Policy 27/28

86 Further Questions Problem Statement Structural Results Bisection Policy 28/28

87 Further Questions If we were allowed to evaluate the function at multiple points in parallel, where would we query the function? Which subintervals do we refine at each time epoch? How much computing power do we allocate to each subinterval? Problem Statement Structural Results Bisection Policy 28/28

88 Further Questions If we were allowed to evaluate the function at multiple points in parallel, where would we query the function? Which subintervals do we refine at each time epoch? How much computing power do we allocate to each subinterval? What if we were only interested in sorting the rankings of the top k zeros? What would a bisection policy look like? Can we effectively compute the optimal policy if the i.i.d. assumption does not hold? Problem Statement Structural Results Bisection Policy 28/28

Coupled Bisection for Root Ordering

Coupled Bisection for Root Ordering Coupled Bisection for Root Ordering Stephen N. Pallone, Peter I. Frazier, Shane G. Henderson School of Operations Research and Information Engineering Cornell University, Ithaca, NY 14853 Abstract We consider

More information

Bayesian Multiple Target Localization

Bayesian Multiple Target Localization Bayesian Multiple Target Localization Purnima Rajan (Johns Hopkins) Weidong Han (Princeton) Raphael Sznitman (Bern) Peter Frazier (Cornell, Uber) Bruno Jedynak (Johns Hopkins, Portland State) Thursday

More information

Bayesian Contextual Multi-armed Bandits

Bayesian Contextual Multi-armed Bandits Bayesian Contextual Multi-armed Bandits Xiaoting Zhao Joint Work with Peter I. Frazier School of Operations Research and Information Engineering Cornell University October 22, 2012 1 / 33 Outline 1 Motivating

More information

Performance Measures for Ranking and Selection Procedures

Performance Measures for Ranking and Selection Procedures Rolf Waeber Performance Measures for Ranking and Selection Procedures 1/23 Performance Measures for Ranking and Selection Procedures Rolf Waeber Peter I. Frazier Shane G. Henderson Operations Research

More information

The Game of Twenty Questions with noisy answers. Applications to Fast face detection, micro-surgical tool tracking and electron microscopy

The Game of Twenty Questions with noisy answers. Applications to Fast face detection, micro-surgical tool tracking and electron microscopy The Game of Twenty Questions with noisy answers. Applications to Fast face detection, micro-surgical tool tracking and electron microscopy Graduate Summer School: Computer Vision July 22 - August 9, 2013

More information

1 Inverse Transform Method and some alternative algorithms

1 Inverse Transform Method and some alternative algorithms Copyright c 2016 by Karl Sigman 1 Inverse Transform Method and some alternative algorithms Assuming our computer can hand us, upon demand, iid copies of rvs that are uniformly distributed on (0, 1), it

More information

MS&E 321 Spring Stochastic Systems June 1, 2013 Prof. Peter W. Glynn Page 1 of 10

MS&E 321 Spring Stochastic Systems June 1, 2013 Prof. Peter W. Glynn Page 1 of 10 MS&E 321 Spring 12-13 Stochastic Systems June 1, 2013 Prof. Peter W. Glynn Page 1 of 10 Section 3: Regenerative Processes Contents 3.1 Regeneration: The Basic Idea............................... 1 3.2

More information

Chapter 2 Solutions of Equations of One Variable

Chapter 2 Solutions of Equations of One Variable Chapter 2 Solutions of Equations of One Variable 2.1 Bisection Method In this chapter we consider one of the most basic problems of numerical approximation, the root-finding problem. This process involves

More information

ADAPTIVE BAYES-OPTIMAL METHODS FOR STOCHASTIC SEARCH WITH APPLICATIONS TO PREFERENCE LEARNING

ADAPTIVE BAYES-OPTIMAL METHODS FOR STOCHASTIC SEARCH WITH APPLICATIONS TO PREFERENCE LEARNING ADAPTIVE BAYES-OPTIMAL METHODS FOR STOCHASTIC SEARCH WITH APPLICATIONS TO PREFERENCE LEARNING A Dissertation Presented to the Faculty of the Graduate School of Cornell University in Partial Fulfillment

More information

Lebesgue Measure on R n

Lebesgue Measure on R n CHAPTER 2 Lebesgue Measure on R n Our goal is to construct a notion of the volume, or Lebesgue measure, of rather general subsets of R n that reduces to the usual volume of elementary geometrical sets

More information

Homework 4, 5, 6 Solutions. > 0, and so a n 0 = n + 1 n = ( n+1 n)( n+1+ n) 1 if n is odd 1/n if n is even diverges.

Homework 4, 5, 6 Solutions. > 0, and so a n 0 = n + 1 n = ( n+1 n)( n+1+ n) 1 if n is odd 1/n if n is even diverges. 2..2(a) lim a n = 0. Homework 4, 5, 6 Solutions Proof. Let ɛ > 0. Then for n n = 2+ 2ɛ we have 2n 3 4+ ɛ 3 > ɛ > 0, so 0 < 2n 3 < ɛ, and thus a n 0 = 2n 3 < ɛ. 2..2(g) lim ( n + n) = 0. Proof. Let ɛ >

More information

Question: My computer only knows how to generate a uniform random variable. How do I generate others? f X (x)dx. f X (s)ds.

Question: My computer only knows how to generate a uniform random variable. How do I generate others? f X (x)dx. f X (s)ds. Simulation Question: My computer only knows how to generate a uniform random variable. How do I generate others?. Continuous Random Variables Recall that a random variable X is continuous if it has a probability

More information

Yi Wang Department of Applied Mathematics, Dalian University of Technology, Dalian , China (Submitted June 2002)

Yi Wang Department of Applied Mathematics, Dalian University of Technology, Dalian , China (Submitted June 2002) SELF-INVERSE SEQUENCES RELATED TO A BINOMIAL INVERSE PAIR Yi Wang Department of Applied Mathematics, Dalian University of Technology, Dalian 116024, China (Submitted June 2002) 1 INTRODUCTION Pairs of

More information

Ordinal optimization - Empirical large deviations rate estimators, and multi-armed bandit methods

Ordinal optimization - Empirical large deviations rate estimators, and multi-armed bandit methods Ordinal optimization - Empirical large deviations rate estimators, and multi-armed bandit methods Sandeep Juneja Tata Institute of Fundamental Research Mumbai, India joint work with Peter Glynn Applied

More information

POISSON PROCESSES 1. THE LAW OF SMALL NUMBERS

POISSON PROCESSES 1. THE LAW OF SMALL NUMBERS POISSON PROCESSES 1. THE LAW OF SMALL NUMBERS 1.1. The Rutherford-Chadwick-Ellis Experiment. About 90 years ago Ernest Rutherford and his collaborators at the Cavendish Laboratory in Cambridge conducted

More information

Lecture 6: Non-stochastic best arm identification

Lecture 6: Non-stochastic best arm identification CSE599i: Online and Adaptive Machine Learning Winter 08 Lecturer: Kevin Jamieson Lecture 6: Non-stochastic best arm identification Scribes: Anran Wang, eibin Li, rian Chan, Shiqing Yu, Zhijin Zhou Disclaimer:

More information

Lecture 5: Loop Invariants and Insertion-sort

Lecture 5: Loop Invariants and Insertion-sort Lecture 5: Loop Invariants and Insertion-sort COMS10007 - Algorithms Dr. Christian Konrad 11.01.2019 Dr. Christian Konrad Lecture 5: Loop Invariants and Insertion-sort 1 / 12 Proofs by Induction Structure

More information

Fundamental Inequalities, Convergence and the Optional Stopping Theorem for Continuous-Time Martingales

Fundamental Inequalities, Convergence and the Optional Stopping Theorem for Continuous-Time Martingales Fundamental Inequalities, Convergence and the Optional Stopping Theorem for Continuous-Time Martingales Prakash Balachandran Department of Mathematics Duke University April 2, 2008 1 Review of Discrete-Time

More information

Chapter 6 Expectation and Conditional Expectation. Lectures Definition 6.1. Two random variables defined on a probability space are said to be

Chapter 6 Expectation and Conditional Expectation. Lectures Definition 6.1. Two random variables defined on a probability space are said to be Chapter 6 Expectation and Conditional Expectation Lectures 24-30 In this chapter, we introduce expected value or the mean of a random variable. First we define expectation for discrete random variables

More information

Computational Oracle Inequalities for Large Scale Model Selection Problems

Computational Oracle Inequalities for Large Scale Model Selection Problems for Large Scale Model Selection Problems University of California at Berkeley Queensland University of Technology ETH Zürich, September 2011 Joint work with Alekh Agarwal, John Duchi and Clément Levrard.

More information

Stability Analysis and Synthesis for Scalar Linear Systems With a Quantized Feedback

Stability Analysis and Synthesis for Scalar Linear Systems With a Quantized Feedback IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL 48, NO 9, SEPTEMBER 2003 1569 Stability Analysis and Synthesis for Scalar Linear Systems With a Quantized Feedback Fabio Fagnani and Sandro Zampieri Abstract

More information

Mathematics for Decision Making: An Introduction. Lecture 8

Mathematics for Decision Making: An Introduction. Lecture 8 Mathematics for Decision Making: An Introduction Lecture 8 Matthias Köppe UC Davis, Mathematics January 29, 2009 8 1 Shortest Paths and Feasible Potentials Feasible Potentials Suppose for all v V, there

More information

MAT1000 ASSIGNMENT 1. a k 3 k. x =

MAT1000 ASSIGNMENT 1. a k 3 k. x = MAT1000 ASSIGNMENT 1 VITALY KUZNETSOV Question 1 (Exercise 2 on page 37). Tne Cantor set C can also be described in terms of ternary expansions. (a) Every number in [0, 1] has a ternary expansion x = a

More information

Lecture 13. Poisson Distribution. Text: A Course in Probability by Weiss 5.5. STAT 225 Introduction to Probability Models February 16, 2014

Lecture 13. Poisson Distribution. Text: A Course in Probability by Weiss 5.5. STAT 225 Introduction to Probability Models February 16, 2014 Lecture 13 Text: A Course in Probability by Weiss 5.5 STAT 225 Introduction to Probability Models February 16, 2014 Whitney Huang Purdue University 13.1 Agenda 1 2 3 13.2 Review So far, we have seen discrete

More information

Lecture 7. µ(x)f(x). When µ is a probability measure, we say µ is a stationary distribution.

Lecture 7. µ(x)f(x). When µ is a probability measure, we say µ is a stationary distribution. Lecture 7 1 Stationary measures of a Markov chain We now study the long time behavior of a Markov Chain: in particular, the existence and uniqueness of stationary measures, and the convergence of the distribution

More information

Reverse mathematics of some topics from algorithmic graph theory

Reverse mathematics of some topics from algorithmic graph theory F U N D A M E N T A MATHEMATICAE 157 (1998) Reverse mathematics of some topics from algorithmic graph theory by Peter G. C l o t e (Chestnut Hill, Mass.) and Jeffry L. H i r s t (Boone, N.C.) Abstract.

More information

Lecture 3 Continuous Random Variable

Lecture 3 Continuous Random Variable Lecture 3 Continuous Random Variable 1 Cumulative Distribution Function Definition Theorem 3.1 For any random variable X, 2 Continuous Random Variable Definition 3 Example Suppose we have a wheel of circumference

More information

Statistics 3858 : Maximum Likelihood Estimators

Statistics 3858 : Maximum Likelihood Estimators Statistics 3858 : Maximum Likelihood Estimators 1 Method of Maximum Likelihood In this method we construct the so called likelihood function, that is L(θ) = L(θ; X 1, X 2,..., X n ) = f n (X 1, X 2,...,

More information

Lecture 9: Dynamics in Load Balancing

Lecture 9: Dynamics in Load Balancing Computational Game Theory Spring Semester, 2003/4 Lecture 9: Dynamics in Load Balancing Lecturer: Yishay Mansour Scribe: Anat Axelrod, Eran Werner 9.1 Lecture Overview In this lecture we consider dynamics

More information

Mathematical Statistics 1 Math A 6330

Mathematical Statistics 1 Math A 6330 Mathematical Statistics 1 Math A 6330 Chapter 2 Transformations and Expectations Mohamed I. Riffi Department of Mathematics Islamic University of Gaza September 14, 2015 Outline 1 Distributions of Functions

More information

Computing Connected Components Given a graph G = (V; E) compute the connected components of G. Connected-Components(G) 1 for each vertex v 2 V [G] 2 d

Computing Connected Components Given a graph G = (V; E) compute the connected components of G. Connected-Components(G) 1 for each vertex v 2 V [G] 2 d Data Structures for Disjoint Sets Maintain a Dynamic collection of disjoint sets. Each set has a unique representative (an arbitrary member of the set). x. Make-Set(x) - Create a new set with one member

More information

1 Approximating area under curves and Riemann sums

1 Approximating area under curves and Riemann sums Professor Jennifer Balakrishnan, jbala@bu.edu What is on today 1 Approximating area under curves and Riemann sums 1 1.1 Riemann sums................................... 1 1.2 Area under the velocity curve..........................

More information

Ordinal Optimization and Multi Armed Bandit Techniques

Ordinal Optimization and Multi Armed Bandit Techniques Ordinal Optimization and Multi Armed Bandit Techniques Sandeep Juneja. with Peter Glynn September 10, 2014 The ordinal optimization problem Determining the best of d alternative designs for a system, on

More information

Modelling the risk process

Modelling the risk process Modelling the risk process Krzysztof Burnecki Hugo Steinhaus Center Wroc law University of Technology www.im.pwr.wroc.pl/ hugo Modelling the risk process 1 Risk process If (Ω, F, P) is a probability space

More information

Lebesgue Measure on R n

Lebesgue Measure on R n 8 CHAPTER 2 Lebesgue Measure on R n Our goal is to construct a notion of the volume, or Lebesgue measure, of rather general subsets of R n that reduces to the usual volume of elementary geometrical sets

More information

Constructing c-ary Perfect Factors

Constructing c-ary Perfect Factors Constructing c-ary Perfect Factors Chris J. Mitchell Computer Science Department Royal Holloway University of London Egham Hill Egham Surrey TW20 0EX England. Tel.: +44 784 443423 Fax: +44 784 443420 Email:

More information

MATH 131A: REAL ANALYSIS (BIG IDEAS)

MATH 131A: REAL ANALYSIS (BIG IDEAS) MATH 131A: REAL ANALYSIS (BIG IDEAS) Theorem 1 (The Triangle Inequality). For all x, y R we have x + y x + y. Proposition 2 (The Archimedean property). For each x R there exists an n N such that n > x.

More information

Mathematics for Engineers. Numerical mathematics

Mathematics for Engineers. Numerical mathematics Mathematics for Engineers Numerical mathematics Integers Determine the largest representable integer with the intmax command. intmax ans = int32 2147483647 2147483647+1 ans = 2.1475e+09 Remark The set

More information

Sequential Selection of Projects

Sequential Selection of Projects Sequential Selection of Projects Kemal Gürsoy Rutgers University, Department of MSIS, New Jersey, USA Fusion Fest October 11, 2014 Outline 1 Introduction Model 2 Necessary Knowledge Sequential Statistics

More information

Algorithm Analysis Recurrence Relation. Chung-Ang University, Jaesung Lee

Algorithm Analysis Recurrence Relation. Chung-Ang University, Jaesung Lee Algorithm Analysis Recurrence Relation Chung-Ang University, Jaesung Lee Recursion 2 Recursion 3 Recursion in Real-world Fibonacci sequence = + Initial conditions: = 0 and = 1. = + = + = + 0, 1, 1, 2,

More information

7 The structure of graphs excluding a topological minor

7 The structure of graphs excluding a topological minor 7 The structure of graphs excluding a topological minor Grohe and Marx [39] proved the following structure theorem for graphs excluding a topological minor: Theorem 7.1 ([39]). For every positive integer

More information

MDP Preliminaries. Nan Jiang. February 10, 2019

MDP Preliminaries. Nan Jiang. February 10, 2019 MDP Preliminaries Nan Jiang February 10, 2019 1 Markov Decision Processes In reinforcement learning, the interactions between the agent and the environment are often described by a Markov Decision Process

More information

1. (Regular) Exponential Family

1. (Regular) Exponential Family 1. (Regular) Exponential Family The density function of a regular exponential family is: [ ] Example. Poisson(θ) [ ] Example. Normal. (both unknown). ) [ ] [ ] [ ] [ ] 2. Theorem (Exponential family &

More information

Chapter 4. Greedy Algorithms. Slides by Kevin Wayne. Copyright 2005 Pearson-Addison Wesley. All rights reserved.

Chapter 4. Greedy Algorithms. Slides by Kevin Wayne. Copyright 2005 Pearson-Addison Wesley. All rights reserved. Chapter 4 Greedy Algorithms Slides by Kevin Wayne. Copyright 2005 Pearson-Addison Wesley. All rights reserved. 1 4.1 Interval Scheduling Interval Scheduling Interval scheduling. Job j starts at s j and

More information

Irredundant generating sets and dimension-like invariants of the finite group arxiv: v2 [math.gr] 13 Dec 2017

Irredundant generating sets and dimension-like invariants of the finite group arxiv: v2 [math.gr] 13 Dec 2017 Irredundant generating sets and dimension-like invariants of the finite group arxiv:1712.03923v2 [math.gr] 13 Dec 2017 Minh Nguyen April 9, 2018 Abstract Whiston proved that the maximum size of an irredundant

More information

CS156: The Calculus of Computation

CS156: The Calculus of Computation Page 1 of 61 CS156: The Calculus of Computation Zohar Manna Winter 2010 Chapter 5: Program Correctness: Mechanics Page 2 of 61 Program A: LinearSearch with function specification @pre 0 l u < a @post rv

More information

Computer Science & Engineering 423/823 Design and Analysis of Algorithms

Computer Science & Engineering 423/823 Design and Analysis of Algorithms Bipartite Matching Computer Science & Engineering 423/823 Design and s Lecture 07 (Chapter 26) Stephen Scott (Adapted from Vinodchandran N. Variyam) 1 / 36 Spring 2010 Bipartite Matching 2 / 36 Can use

More information

Models of collective inference

Models of collective inference Models of collective inference Laurent Massoulié (Microsoft Research-Inria Joint Centre) Mesrob I. Ohannessian (University of California, San Diego) Alexandre Proutière (KTH Royal Institute of Technology)

More information

1 Complete Statistics

1 Complete Statistics Complete Statistics February 4, 2016 Debdeep Pati 1 Complete Statistics Suppose X P θ, θ Θ. Let (X (1),..., X (n) ) denote the order statistics. Definition 1. A statistic T = T (X) is complete if E θ g(t

More information

3.1 Interpolation and the Lagrange Polynomial

3.1 Interpolation and the Lagrange Polynomial MATH 4073 Chapter 3 Interpolation and Polynomial Approximation Fall 2003 1 Consider a sample x x 0 x 1 x n y y 0 y 1 y n. Can we get a function out of discrete data above that gives a reasonable estimate

More information

ORBIT-HOMOGENEITY. Abstract

ORBIT-HOMOGENEITY. Abstract Submitted exclusively to the London Mathematical Society DOI: 10.1112/S0000000000000000 ORBIT-HOMOGENEITY PETER J. CAMERON and ALEXANDER W. DENT Abstract We introduce the concept of orbit-homogeneity of

More information

Exercises for Unit VI (Infinite constructions in set theory)

Exercises for Unit VI (Infinite constructions in set theory) Exercises for Unit VI (Infinite constructions in set theory) VI.1 : Indexed families and set theoretic operations (Halmos, 4, 8 9; Lipschutz, 5.3 5.4) Lipschutz : 5.3 5.6, 5.29 5.32, 9.14 1. Generalize

More information

MEASURABLE PARTITIONS, DISINTEGRATION AND CONDITIONAL MEASURES

MEASURABLE PARTITIONS, DISINTEGRATION AND CONDITIONAL MEASURES MEASURABLE PARTITIONS, DISINTEGRATION AND CONDITIONAL MEASURES BRUNO SANTIAGO Abstract. In this short note we review Rokhlin Desintegration Theorem and give some applications. 1. Introduction Consider

More information

MA 323 Geometric Modelling Course Notes: Day 12 de Casteljau s Algorithm and Subdivision

MA 323 Geometric Modelling Course Notes: Day 12 de Casteljau s Algorithm and Subdivision MA 323 Geometric Modelling Course Notes: Day 12 de Casteljau s Algorithm and Subdivision David L. Finn Yesterday, we introduced barycentric coordinates and de Casteljau s algorithm. Today, we want to go

More information

On the Complexity of Computing an Equilibrium in Combinatorial Auctions

On the Complexity of Computing an Equilibrium in Combinatorial Auctions On the Complexity of Computing an Equilibrium in Combinatorial Auctions Shahar Dobzinski Hu Fu Robert Kleinberg April 8, 2014 Abstract We study combinatorial auctions where each item is sold separately

More information

Cover Page. The handle holds various files of this Leiden University dissertation

Cover Page. The handle   holds various files of this Leiden University dissertation Cover Page The handle http://hdlhandlenet/1887/39637 holds various files of this Leiden University dissertation Author: Smit, Laurens Title: Steady-state analysis of large scale systems : the successive

More information

CTL Model Checking. Prof. P.H. Schmitt. Formal Systems II. Institut für Theoretische Informatik Fakultät für Informatik Universität Karlsruhe (TH)

CTL Model Checking. Prof. P.H. Schmitt. Formal Systems II. Institut für Theoretische Informatik Fakultät für Informatik Universität Karlsruhe (TH) CTL Model Checking Prof. P.H. Schmitt Institut für Theoretische Informatik Fakultät für Informatik Universität Karlsruhe (TH) Formal Systems II Prof. P.H. Schmitt CTLMC Summer 2009 1 / 26 Fixed Point Theory

More information

c 2013 Society for Industrial and Applied Mathematics

c 2013 Society for Industrial and Applied Mathematics SIAM J. CONTROL OPTIM. Vol. 51, No. 3, pp. 2261 2279 c 2013 Society for Industrial and Applied Mathematics BISECTION SEARCH WITH NOISY RESPONSES ROLF WAEBER, PETER I. FRAZIER, AND SHANE G. HENDERSON Abstract.

More information

3.3 Accumulation Sequences

3.3 Accumulation Sequences 3.3. ACCUMULATION SEQUENCES 25 3.3 Accumulation Sequences Overview. One of the most important mathematical ideas in calculus is that of an accumulation of change for physical quantities. As we have been

More information

Transformations and Expectations

Transformations and Expectations Transformations and Expectations 1 Distributions of Functions of a Random Variable If is a random variable with cdf F (x), then any function of, say g(), is also a random variable. Sine Y = g() is a function

More information

Notes on Iterated Expectations Stephen Morris February 2002

Notes on Iterated Expectations Stephen Morris February 2002 Notes on Iterated Expectations Stephen Morris February 2002 1. Introduction Consider the following sequence of numbers. Individual 1's expectation of random variable X; individual 2's expectation of individual

More information

THE INVERSE FUNCTION THEOREM

THE INVERSE FUNCTION THEOREM THE INVERSE FUNCTION THEOREM W. PATRICK HOOPER The implicit function theorem is the following result: Theorem 1. Let f be a C 1 function from a neighborhood of a point a R n into R n. Suppose A = Df(a)

More information

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Electrical Engineering and Computer Science

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Electrical Engineering and Computer Science MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Electrical Engineering and Computer Science 6.262 Discrete Stochastic Processes Midterm Quiz April 6, 2010 There are 5 questions, each with several parts.

More information

Notes on arithmetic. 1. Representation in base B

Notes on arithmetic. 1. Representation in base B Notes on arithmetic The Babylonians that is to say, the people that inhabited what is now southern Iraq for reasons not entirely clear to us, ued base 60 in scientific calculation. This offers us an excuse

More information

Scheduling Periodic Real-Time Tasks on Uniprocessor Systems. LS 12, TU Dortmund

Scheduling Periodic Real-Time Tasks on Uniprocessor Systems. LS 12, TU Dortmund Scheduling Periodic Real-Time Tasks on Uniprocessor Systems Prof. Dr. Jian-Jia Chen LS 12, TU Dortmund 08, Dec., 2015 Prof. Dr. Jian-Jia Chen (LS 12, TU Dortmund) 1 / 38 Periodic Control System Pseudo-code

More information

A Subrecursive Refinement of FTA. of the Fundamental Theorem of Algebra

A Subrecursive Refinement of FTA. of the Fundamental Theorem of Algebra A Subrecursive Refinement of the Fundamental Theorem of Algebra P. Peshev D. Skordev Faculty of Mathematics and Computer Science University of Sofia Computability in Europe, 2006 Outline Introduction 1

More information

0.1 Minimizing Finite State Machines

0.1 Minimizing Finite State Machines .. MINIMIZING FINITE STATE MACHINES. Minimizing Finite State Machines Here we discuss the problem of minimizing the number of states of a DFA. We will show that for any regular language L, thereisaunique

More information

Bayesian Multiple Target Localization

Bayesian Multiple Target Localization Purnima Rajan Department of Computer Science, Johns Hopkins University Weidong Han Department of Operations Research and Financial Engineering, Princeton University PURNIMA@CS.JHU.EDU WHAN@INCETON.EDU

More information

Stat 315: HW #6. Fall Due: Wednesday, October 10, 2018

Stat 315: HW #6. Fall Due: Wednesday, October 10, 2018 Stat 315: HW #6 Fall 018 Due: Wednesday, October 10, 018 Updated: Monday, October 8 for misprints. 1. An airport shuttle route includes two intersections with traffic lights. Let i be the number of lights

More information

The Contraction Method on C([0, 1]) and Donsker s Theorem

The Contraction Method on C([0, 1]) and Donsker s Theorem The Contraction Method on C([0, 1]) and Donsker s Theorem Henning Sulzbach J. W. Goethe-Universität Frankfurt a. M. YEP VII Probability, random trees and algorithms Eindhoven, March 12, 2010 joint work

More information

Algorithm Design and Analysis

Algorithm Design and Analysis Algorithm Design and Analysis LECTURE 9 Divide and Conquer Merge sort Counting Inversions Binary Search Exponentiation Solving Recurrences Recursion Tree Method Master Theorem Sofya Raskhodnikova S. Raskhodnikova;

More information

Multi-Attribute Bayesian Optimization under Utility Uncertainty

Multi-Attribute Bayesian Optimization under Utility Uncertainty Multi-Attribute Bayesian Optimization under Utility Uncertainty Raul Astudillo Cornell University Ithaca, NY 14853 ra598@cornell.edu Peter I. Frazier Cornell University Ithaca, NY 14853 pf98@cornell.edu

More information

Things to remember when learning probability distributions:

Things to remember when learning probability distributions: SPECIAL DISTRIBUTIONS Some distributions are special because they are useful They include: Poisson, exponential, Normal (Gaussian), Gamma, geometric, negative binomial, Binomial and hypergeometric distributions

More information

Joint-optimal Probing and Scheduling in Wireless Systems

Joint-optimal Probing and Scheduling in Wireless Systems Joint-optimal Probing and Scheduling in Wireless Systems Prasanna Chaporkar and Alexandre Proutiere Abstract Consider a wireless system where a sender can transmit data to various users with independent

More information

Lecture 10. Riemann-Stieltjes Integration

Lecture 10. Riemann-Stieltjes Integration Lecture 10 Riemann-Stieltjes Integration In this section we will develop the basic definitions and some of the properties of the Riemann-Stieltjes Integral. The development will follow that of the Darboux

More information

Algorithms for pattern involvement in permutations

Algorithms for pattern involvement in permutations Algorithms for pattern involvement in permutations M. H. Albert Department of Computer Science R. E. L. Aldred Department of Mathematics and Statistics M. D. Atkinson Department of Computer Science D.

More information

CSE 421 Greedy Algorithms / Interval Scheduling

CSE 421 Greedy Algorithms / Interval Scheduling CSE 421 Greedy Algorithms / Interval Scheduling Yin Tat Lee 1 Interval Scheduling Job j starts at s(j) and finishes at f(j). Two jobs compatible if they don t overlap. Goal: find maximum subset of mutually

More information

Lecture 2: A Las Vegas Algorithm for finding the closest pair of points in the plane

Lecture 2: A Las Vegas Algorithm for finding the closest pair of points in the plane Randomized Algorithms Lecture 2: A Las Vegas Algorithm for finding the closest pair of points in the plane Sotiris Nikoletseas Professor CEID - ETY Course 2017-2018 Sotiris Nikoletseas, Professor Randomized

More information

An Online Algorithm for Maximizing Submodular Functions

An Online Algorithm for Maximizing Submodular Functions An Online Algorithm for Maximizing Submodular Functions Matthew Streeter December 20, 2007 CMU-CS-07-171 Daniel Golovin School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 This research

More information

Process-Based Risk Measures for Observable and Partially Observable Discrete-Time Controlled Systems

Process-Based Risk Measures for Observable and Partially Observable Discrete-Time Controlled Systems Process-Based Risk Measures for Observable and Partially Observable Discrete-Time Controlled Systems Jingnan Fan Andrzej Ruszczyński November 5, 2014; revised April 15, 2015 Abstract For controlled discrete-time

More information

Lecture 20 FUNDAMENTAL Theorem of Finitely Generated Abelian Groups (FTFGAG)

Lecture 20 FUNDAMENTAL Theorem of Finitely Generated Abelian Groups (FTFGAG) Lecture 20 FUNDAMENTAL Theorem of Finitely Generated Abelian Groups (FTFGAG) Warm up: 1. Let n 1500. Find all sequences n 1 n 2... n s 2 satisfying n i 1 and n 1 n s n (where s can vary from sequence to

More information

Application: Bucket Sort

Application: Bucket Sort 5.2.2. Application: Bucket Sort Bucket sort breaks the log) lower bound for standard comparison-based sorting, under certain assumptions on the input We want to sort a set of =2 integers chosen I+U@R from

More information

Quadratic Interval Refinement

Quadratic Interval Refinement (QIR) Analysis of Seminar on Computational Geometry and Geometric Computing Outline Introduction (QIR) Analysis of 1 Introduction in Short Main Goal of this Work 2 3 (QIR) Bisection Method Algorithm 4

More information

Theme 1: Solving Nonlinear Equations

Theme 1: Solving Nonlinear Equations Theme 1: Solving Nonlinear Equations Prof. Mapundi K. Banda (Tuks) WTW263 Semester II 1 / 22 Sources of Errors Finite decimal representation (Rounding): Finite decimal representation will be used to represent

More information

COMPUTING OPTIMAL SEQUENTIAL ALLOCATION RULES IN CLINICAL TRIALS* Michael N. Katehakis. State University of New York at Stony Brook. and.

COMPUTING OPTIMAL SEQUENTIAL ALLOCATION RULES IN CLINICAL TRIALS* Michael N. Katehakis. State University of New York at Stony Brook. and. COMPUTING OPTIMAL SEQUENTIAL ALLOCATION RULES IN CLINICAL TRIALS* Michael N. Katehakis State University of New York at Stony Brook and Cyrus Derman Columbia University The problem of assigning one of several

More information

Termination of LCTRSs

Termination of LCTRSs Termination of LCTRSs Cynthia Kop 1 1 Department of Computer Science, University of Innsbruck Technikerstraße 21a, 6020 Innsbruck, Austria Cynthia.Kop@uibk.ac.at Abstract Logically Constrained Term Rewriting

More information

C.7. Numerical series. Pag. 147 Proof of the converging criteria for series. Theorem 5.29 (Comparison test) Let a k and b k be positive-term series

C.7. Numerical series. Pag. 147 Proof of the converging criteria for series. Theorem 5.29 (Comparison test) Let a k and b k be positive-term series C.7 Numerical series Pag. 147 Proof of the converging criteria for series Theorem 5.29 (Comparison test) Let and be positive-term series such that 0, for any k 0. i) If the series converges, then also

More information

Randomized Load Balancing:The Power of 2 Choices

Randomized Load Balancing:The Power of 2 Choices Randomized Load Balancing: The Power of 2 Choices June 3, 2010 Balls and Bins Problem We have m balls that are thrown into n bins, the location of each ball chosen independently and uniformly at random

More information

Ch3. Generating Random Variates with Non-Uniform Distributions

Ch3. Generating Random Variates with Non-Uniform Distributions ST4231, Semester I, 2003-2004 Ch3. Generating Random Variates with Non-Uniform Distributions This chapter mainly focuses on methods for generating non-uniform random numbers based on the built-in standard

More information

Polyhedral Approaches to Online Bipartite Matching

Polyhedral Approaches to Online Bipartite Matching Polyhedral Approaches to Online Bipartite Matching Alejandro Toriello joint with Alfredo Torrico, Shabbir Ahmed Stewart School of Industrial and Systems Engineering Georgia Institute of Technology Industrial

More information

Chapter 1. Introduction

Chapter 1. Introduction Chapter 1 Introduction 1.1 Areas Modern integration theory is the culmination of centuries of refinements and extensions of ideas dating back to the Greeks. It evolved from the ancient problem of calculating

More information

Stochastic Dynamic Programming: The One Sector Growth Model

Stochastic Dynamic Programming: The One Sector Growth Model Stochastic Dynamic Programming: The One Sector Growth Model Esteban Rossi-Hansberg Princeton University March 26, 2012 Esteban Rossi-Hansberg () Stochastic Dynamic Programming March 26, 2012 1 / 31 References

More information

Standard forms for writing numbers

Standard forms for writing numbers Standard forms for writing numbers In order to relate the abstract mathematical descriptions of familiar number systems to the everyday descriptions of numbers by decimal expansions and similar means,

More information

Intelligent Data Analysis. Principal Component Analysis. School of Computer Science University of Birmingham

Intelligent Data Analysis. Principal Component Analysis. School of Computer Science University of Birmingham Intelligent Data Analysis Principal Component Analysis Peter Tiňo School of Computer Science University of Birmingham Discovering low-dimensional spatial layout in higher dimensional spaces - 1-D/3-D example

More information

Math212a1413 The Lebesgue integral.

Math212a1413 The Lebesgue integral. Math212a1413 The Lebesgue integral. October 28, 2014 Simple functions. In what follows, (X, F, m) is a space with a σ-field of sets, and m a measure on F. The purpose of today s lecture is to develop the

More information

The L Machines are very high-level, in two senses:

The L Machines are very high-level, in two senses: What is a Computer? State of the machine. CMPSCI 630: Programming Languages An Abstract Machine for Control Spring 2009 (with thanks to Robert Harper) Internal registers, memory, etc. Initial and final

More information

ON THE OPTIMALITY OF AN INDEX RULE IN MULTICHANNEL ALLOCATION FOR SINGLE-HOP MOBILE NETWORKS WITH MULTIPLE SERVICE CLASSES*

ON THE OPTIMALITY OF AN INDEX RULE IN MULTICHANNEL ALLOCATION FOR SINGLE-HOP MOBILE NETWORKS WITH MULTIPLE SERVICE CLASSES* Probability in the Engineering and Informational Sciences, 14, 2000, 259 297+ Printed in the U+S+A+ ON THE OPTIMALITY OF AN INDEX RULE IN MULTICHANNEL ALLOCATION FOR SINGLE-HOP MOBILE NETWORKS WITH MULTIPLE

More information

Iowa State University. Instructor: Alex Roitershtein Summer Exam #1. Solutions. x u = 2 x v

Iowa State University. Instructor: Alex Roitershtein Summer Exam #1. Solutions. x u = 2 x v Math 501 Iowa State University Introduction to Real Analysis Department of Mathematics Instructor: Alex Roitershtein Summer 015 Exam #1 Solutions This is a take-home examination. The exam includes 8 questions.

More information

A better lower bound for quantum algorithms searching an ordered list

A better lower bound for quantum algorithms searching an ordered list A better lower bound for quantum algorithms searching an ordered list Andris Ambainis Computer Science Division University of California Berkeley, CA 94720, e-mail: ambainis@cs.berkeley.edu Abstract We

More information