The Mean Median Map Marc Chamberland

Slides:



Advertisements
Similar presentations
Lower Envelopes Yuval Suede. What’s on the menu? Lower Envelopes Davenport-Schinzel Sequences Trivial upper bound Super linear complexity* Tight upper.
Advertisements

Continuing Abstract Interpretation We have seen: 1.How to compile abstract syntax trees into control-flow graphs 2.Lattices, as structures that describe.
Boyce/DiPrima 9th ed, Ch 2.8: The Existence and Uniqueness Theorem Elementary Differential Equations and Boundary Value Problems, 9th edition, by William.
12.5: Absolute Maxima and Minima. Finding the absolute maximum or minimum value of a function is one of the most important uses of the derivative. For.
Chapter 3 Application of Derivatives
Congestion Games with Player- Specific Payoff Functions Igal Milchtaich, Department of Mathematics, The Hebrew University of Jerusalem, 1993 Presentation.
Calculus and Elementary Analysis 1.2 Supremum and Infimum Integers, Rational Numbers and Real Numbers Completeness of Real Numbers Supremum and Infimum.
Copyright © Cengage Learning. All rights reserved.
Mathematics By: Damarlyn Chivers Equations Definitions An equation is two things are the same, using mathematics symbols.
Extrema on an interval (3.1) November 15th, 2012.
MTH374: Optimization For Master of Mathematics By Dr. M. Fazeel Anwar Assistant Professor Department of Mathematics, CIIT Islamabad 1.
Monotone Sequences Objective: To define a Monotone Sequence and determine whether it converges or not.
Copyright © 2011 Pearson, Inc. 1.2 Functions and Their Properties.
Instructor: Prof.Dr.Sahand Daneshvar Presented by: Seyed Iman Taheri Student number: Non linear Optimization Spring EASTERN MEDITERRANEAN.
Chapter 3 Application of Derivatives 3.1 Extreme Values of Functions Absolute maxima or minima are also referred to as global maxima or minima.
Order Statistics The ith order statistic in a set of n elements is the ith smallest element The minimum is thus the 1st order statistic The maximum is.
Section 4.1 Maximum and Minimum Values Applications of Differentiation.
1 ECE-517 Reinforcement Learning in Artificial Intelligence Lecture 7: Finite Horizon MDPs, Dynamic Programming Dr. Itamar Arel College of Engineering.
The Selection Problem. 2 Median and Order Statistics In this section, we will study algorithms for finding the i th smallest element in a set of n elements.
Index FAQ Functions, properties. elementary functions and their inverses 2. előadás.
MAT 213 Brief Calculus Section 4.2 Relative and Absolute Extreme Points.
Applications of Differentiation Calculus Chapter 3.
Calculus and Analytical Geometry Lecture # 13 MTH 104.
Simple nonlinear systems. Nonlinear growth of bugs and the logistic map x i+1 =μx i (1-x i )
Introduction to Optimization
1 What happens to the location estimator if we minimize with a power other that 2? Robert J. Blodgett Statistic Seminar - March 13, 2008.
3.6 First Passage Time Distribution
Theorem Suppose {a n } is non-decreasing and bounded above by a number A. Then {a n } converges to some finite limit a, with a  A. Suppose {b n } is non-increasing.
SEQUENCES A function whose domain is the set of all integers greater than or equal to some integer n 0 is called a sequence. Usually the initial number.
Copyright © Cengage Learning. All rights reserved.
DFA foundations Simone Campanoni
Sequences of numbers appear in diverse situations. If you divide a cake in half, and then divide the remaining half in half, and continue dividing in.
Advanced Mathematics D. Chapter Four The Derivatives in Graphing and Application.
Chapter 3 The Real Numbers.
EMGT 6412/MATH 6665 Mathematical Programming Spring 2016
This is a geometric sequence with explicit function:
Quadratic Functions and Models
Precalculus V. J. Motto Polynomial Functions – Part One
Linear Programming Revised Simplex Method, Duality of LP problems and Sensitivity analysis D Nagesh Kumar, IISc Optimization Methods: M3L5.
Prüfer code algorithm Algorithm (Prüfer code)
Copyright © Cengage Learning. All rights reserved.
Sampling Distributions
Gray Code Can you find an ordering of all the n-bit strings in such a way that two consecutive n-bit strings differed by only one bit? This is called the.
Sampling Distributions
Pigeonhole Principle and Ramsey’s Theorem
Given f(x)= x4 –22x3 +39x2 +14x+120 , answer the following questions:
Objectives for Section 12.5 Absolute Maxima and Minima
Proving Existentials A proof of a statement of the form x P(x) is called an existence proof. If the proof demonstrates how to actually find or construct.
Chapter 12 Graphing and Optimization
3.3 Increasing and Decreasing Functions and the First Derivative Test
Section 4.3 Optimization.
COUNTING AND PROBABILITY
Concurrent Models of Computation
Counting Statistics and Error Prediction
Copyright © Cengage Learning. All rights reserved.
Derivatives in Action Chapter 7.
Chapter 4 Sequences.
Discrete Mathematics 7th edition, 2009
Foundations of Discrete Mathematics
The Selection Problem.
3-1 Extreme Values of Functions.
4.2 Critical Points, Local Maxima and Local Minima
Part 7. Phantoms: Legal States That Cannot Be Constructed
Objectives Graphing Basic Polynomial Functions
Totally Asynchronous Iterative Algorithms
Chapter 12 Graphing and Optimization
Maximum and Minimum Values
Chapter 4 Graphing and Optimization
Presentation transcript:

The Mean Median Map Marc Chamberland Department of Mathematics and Statistics Grinnell College Collaborator: Mario Martelli, Claremont McKenna College

What is the map? Start with the non-empty set Generate the sequence by The Mean-Median Map As usual,

Examples We use bold to indicate that the sequence simply repeats that number thereafter. {5} continues as {5} {1,2,3} continues as {2} {0,2,3} continues as { 3, 9/2,11/2, 3} {0,4,5,6} continues as {15/2, 15/2, 17/2, 19/2, 51/4, 57/4, 15/2} {0,4,5} continues as { 7, 13/2 15/2, 41/4, 47/4, 35/4, 37/4, 39/4, 41/4, 125/8, 135/8, 25/2, 13, 27/2, 14, 29/2, 15, 41/4} Terminating Conjecture: Every initial point will become fixed in a finite number of steps.

Medians Theorem: The sequence of medians is montonic. {5,5} continues as {5} {1,2,3} continues as {2} {0,2,3} continues as { 3, 9/2,11/2, 3} medians: 2, 5/2, 3 {0,4,5,6} continues as {15/2, 15/2, 17/2, 19/2, 51/4, 57/4, 15/2} medians: 9/2, 5, 11/2, 6, 27/4, 15/2 {0,4,5} continues as { 7, 13/2 15/2, 41/4, 47/4, 35/4, 37/4, 39/4, 41/4, 125/8, 135/8, 25/2, 13, 27/2, 14, 29/2, 15, 41/4} medians: 4, 9/2, 5, 23/4, 13/2, 27/4, 7, 29/4, 15/2, 65/8, 35/4, 9, 37/4, 19/2, 39/4, 10, 41/4 Theorem: The sequence of medians is montonic. Conjecture: The median occurs exactly twice before it settles permanently.

Properties with three initial points Let M( a,b,c ) be the limiting median of the initial set {a,b,c} (if it exists). Then scaling and translation holds: Let m(x) = M(0,x,1), so consider only 0<x<1. m(1-x) = 1 – m(x), so consider only 1/2<x<1 (medians increase). so consider only 1/2<x<2/3.

The map m on 1/2<x<2/3 Are you ready for something scary?

Steps to Convergence median Steps needed until convergence. We believe this function is unbounded.

Maxima along the Way median Maximum in orbit

Near x=1/2 Theorem: for |e|<4/333. Proof: The next iterates take the form 1/2 + re where r = 3, 6, 8, 13.5, 16.5, 15, …, 117.5625, …, 41.625. QED All of these values of m are less than 1, and there are 70 iterates. For larger values of e, the median reaches 1, so the limiting median must be at least 1. We found that achieves a limiting median of 1 for 22 values of a.

General Form of m An experimental study suggests: m is affine almost everywhere all rational points are either local minima or in affine segments (i.e. no local maxima or monotonic corners) on either side of local minima, the # of steps to convergence is greater and the same on each side. For example, experimental evidence near the local minimum at x=2/3 suggests (for sufficiently small positive e) Steps: 4 60 60

Conclusion Questions: The function m is sensitive: m(0.6) = 2.4 steps = 32 m(0.600001) = 2.6353166429240703582763671875 steps=12488 Using a CAS (like Maple) is imperative! Questions: Does m have an upper bound? Record: m(0.843) = 4.64546875 Is m continuous? Is it almost always affine? Can one characterize the extrema? Can a rational point be a local maximum? What if one starts with more than 3 points?