Exercise 3: Random Numbers and Monte Carlo

Size: px
Start display at page:

Download "Exercise 3: Random Numbers and Monte Carlo"

Transcription

1 Exercise : Random Numbers and Monte Carlo J. F. Ider Chitham Astrophysics Group, H.H. Wills Physics Laboratory, University of Bristol, Tyndall Avenue, Bristol BS8 TL Accepted March. Received March ; in original form March ABSTRACT This exercise address the use of random numbers in Monte Carlo techniques. These are often the fastest or most straightforward way of analysing complex problems in computational analysis and physical simulations. The degree of success is heavily reliant on the availability and reliability number generators. INTRODUCTION Most Monte Carlo sampling or integration techniques assume a random number generator; a routine that returns a random number drawn from a specified probability distribution. Consider the generation of uniform and statistically independent values x gen, on the open interval (, ). Effective routines should incorporate the following properties: (i) Unpredictability: if the last N values returned by the routine are known, the preceding returned values should not be easily predicted. (ii) Correct coverage: the frequency with which any x gen is generated must be proportional to the underlying probability density function (PDF) P (x). (iii) Uniqueness: True randomness is seen in many aspects of physics, for example constructing a circuit that measures the Johnson noise in a resistor, and using the measurement to generate a value would give a uniquely genuine random number, however this is idealisation is difficult to simulate computationally. Many algorithms are in fact pseudo-random number generators (PRNGs) which are unpredictable, but reproducible. These routines can be reset and the exact same sequence of random numbers can be yielded []. True randomness can however be approximated by seeding the generating algorithm with time to eliminate periods of sequential repetition. This can provide uniquely ordered values as the exact moment the generating function is called within the source code is effectively random to the precision of a computational clock.. Central Limit Theorem In probability theory, the central limit theorem (CLT) states that, the mean of a sufficiently large number of mutually independent random variables, will be approximately normally distributed (in the limit that the number of independent events tending to infinity) providing they have a well-defined expectation value and variance []. A Poisson distribution describing the probability of N discrete events occurring tends to a continuous function known as a normal distribution: lim P (n, µ) N (n, µ, σ) () as shown in Figure. n. [ ] µ n e µ lim n n! σ /σ π e (n µ) () P(n) n Figure. Poisson probabilities for means of µ = {.5, }, note orgin of gaussian resemblence emerging at N = independent events. Frequent observations of discrete distributions tending to relatively continuous ones at large numbers iterations can be seen throughout this exercise according to CLT. TASK A uniform random distribution, P (θ) which generates angles from ( < θ gen < ) must be converted to a non-uniform one, proportional to the sine of the transformed angle P (θ ) sin(θ ) with ( < θ gen < π). Providing both distributions are normalized to unit area () (N.B required factor of / ) conversion is possible via analytical inversion and accept-reject routines using contrasting methods of coordinate transformation. θ gen sin(θ )dθ = θgen n P (θ)dθ () θ gen = Q(θ req) θ req = Q (θ gen) (). Inverse Transform The inversion method is exact when an explicit form of Q is known and most convenient when the inverse function of the indefinite integral of P can be manually calculated. This is the case for some common functions such as sin(θ) (). Very commonly an analytic form for Q is unknown or too complex to work with, so

2 J. F. Ider Chitham Table. Comparative summary of goodness of fit analysis Number Of Random Numbers, N χ inv χ a/r that obtaining an inverse analogously is impractical. θ gen = cos(θ gen). Acceptance-Rejection θ gen = arccos( θ gen) Suppose that for any given value of θ, we can enclose P (θ) entirely inside a shape which is π times R(θ), a uniform random distribution with ( < θ gen < ). To generate P (θ), θ gen is generated according to R(θ) then P (θ) is calculated along with the height of the envelope; another random angle, ( < ϕ < ) is generated and then tested if ϕπr(θ) P (θ) to determine whether or not the angle will contribute to the distribution. If so, accept θ; if not reject until points populate the entire area πr(θ) in a smooth manner []. () P (θ ) P (θ ) Inverse Reject-Accept fiti fit ra π/ π/ π/ π θ (rad) COMPARATIVE ANALYSIS Each technique is successful to a consistently similar extent, with the fitted curves being visually indistinguishable from each other at every logarithmic interval of N shown in Figure.. This is confirmed via analysis of fit quality in terms of Pearson s χ whereby a more reliable (mean) value is deduced by averaging over randomly generated P (θ ) distributions for each value of N, a sample of results is summarised in Table.. Despite this each χ will still change slightly due to random variation making comparative analysis difficult although the general trend for both appeared to be χ as N as expected. The goodness of fit however, is of varying quality across each bin in the distribution as the angles in each bin will be normally distributed within their respective bins and the standard error of the mean will vary across the distribution as shown in Figure.. This was further investigated by calculating the deviation from the ideal fit χ as a function of N as illustrated in Figure. over the range 5 N. The deviation from χ = for the analytical inversion transformation method appeared relatively superior on average, indicated by the greater degree of clustering approaching the horizontal axis. This was unexpected however can be explained via the contrasting computational times required to generate N random angles (Figure. 5) as faster algorithms will have a higher degree of similarity between consecutively (time) generated seeds. The linear fit present in the computational time analysis suggests an exponential increase in cumulative generation time as N increases, the relative gradients indicate the generation rate via the reject-accept method is slower. This is to be expected as not every generated angle will contribute to the distribution. The ratio of the total number of generations to the number that successfully contribute to the distribution, η should increasingly resemble the ratio of integrated areas under the distribution and the total, η = /π as the number of random numbers is increased (5). η = N N i N dis i lim N η η = η = π (5) P (θ ). Inverse Reject-Accept fiti fit ra π/ π/ π/ π θ (rad). Inverse Reject-Accept fiti fit ra π/ π/ π/ π θ (rad) Figure. Histograms of P (θ ) (normalized frequency) for inverse transform (red) and reject-accept (blue) methods with fitted sine profiles for N = {,,5,6} (goodness of fit summarised in Table..) The deviation between discrete and integral solutions η = η η, was investigated via a percentage difference as a function of N as summarised by Figure. 6. The difference between each run emphasises the true randomness of the generator whilst the tendency towards continuity outlines the basis of the central limit theorem as mentioned in.. Despite more in depth analysis suggesting the analytical transform method is superior in terms of precision, computational time and number of iterations, this is to be expected considering the simplistic form of P (θ ). If the PDF is more complex inversion may not be a viable option meaning each method is appropriate for dif-

3 Exercise : Random Numbers and Monte Carlo. Standard Error. e-5 η / η (%) e Bin N Figure. Standard error σ N = σ/ N, accross each bin in a N = 5 analytical transformation distribution as shown in Figure.. χ Inverse Transform Reject-Accept Figure. Deviation of mean χ (each averaged over random P (θ ) distributions) from its ideal value as a function of N for for inverse transform (red) and reject-accept (blue) methods. N Figure 6. Percentage difference between discrete and integral ratios of areas, η of P (θ ) for the accept reject method (5 random runs). Note logarithmic axis and approximate exponential drop off as N. P (x) x N= N= N= 5 N= 6 Figure 7. Histograms of Gaussian transformed distribution (via reject accept see ) P (x ) (mean µ =.5, standard deviation σ =.5) for N = {,,5,6}, N bins =. Computational Time (ms).. Inverse Reject-Accept ferent situations (reject-accept proved more applicable for a Gaussian PDF, Figure. 7). Occasionally a combination of the two is required for example, how would you generate random numbers from the distribution: f(θ) = + cos(θ)/ θ. Accept-reject method is not possible due to the asymptote at θ = (see Figure. 8) nor is the transform method because the integral is not easily calculated, a hybrid of the two is required log (N) Figure 5. Computational time to generate log N random angles within the normalized distribution P (θ ) for inverse transform (red) and rejectaccept (blue), note logarmic time axis, note more significant contrast at large N and approximatley constant gradient at N. PHYSICS PROBLEM A beam of unstable, unpolarised nuclei are initially injected into a m cavity perpendicular to an electromagnetic calorimeter array at ms (relativistic effects can be considered negligible as β << ). The radioactive decay rate is well represented by an exponential random probability distribution with a mean lifetime of τ = 5µs (mean path length of l =.m) producing isotropic gamma ray emission. In terms of spherical polar coordinates with respect to the x-y plane of the detector array and the z-axis defined by the path of the nuclei, this translates to a random distribution

4 J. F. Ider Chitham Original Smeared f(θ) θ (rads) Figure 8. f(θ) = + cos(θ)/ θ of azimuth ϕ, and polar θ angles (see.). For a sample of N nuclei at t =, the predicted number decaying in the interval from t to t + dt is given by dn(t) = N τ [exp( t/τ)]dt. To more realistically simulate an electromagnetic calorimeter, a quantum efficiency ɛ, was implemented via the inclusion Bernoulli random distribution (6). This is either or depending on the relative success or failure of photon detection on incidence as determined by a the sample mean probability p (equivalent to the quantum efficiency of each pixel). { f(count, p = ɛ) = p count = p count = This allows the number of detected gamma rays dn γ in the same time interval to be predicted (7), which can be alternately written in terms of displacement (t z, τ l).. Gaussian Smearing (6) dn γ(t) = ɛn τ e t/τ dt (7) When a photon is incident on a (finitely sized) pixel its exact location of incidence across the pixel can not be determined with certainty, therefore an additional probability distribution is required to simulate this limitation of precision. This was achieved via Gaussian smearing with the resolution of the detector array, x res = cm and y res = cm taken as the respective one dimensional full width at half maximum (FWHM) components of the two dimensional Gaussian spread. Combining the FWHM and the mathematical definition of a Gaussian distribution () an expression for the standard deviation was derived, σ = F W HM/ ln. The original Cartesian coordinates of the incident γ-ray (8) were altered via the addition of a Gaussian GSL RNG centred at µ =. x s = x + N x(x, σ x, µ x = ) y s = y + N y(y, σ y, µ y = ) A preliminary investigation was carried out to ensure that the influence of Gaussian smearing was fully functional. Figure. 9 shows a definite alteration in apparent position post smearing, pairs of slightly differing points can easily be identified around the periphery although difficult to distinguish in the central vicinity. There appears to be more significant smearing in the y direction, this is to be expected to be on average a factor more than in the x direction due to the respective ratio of FWHM s. The sample data (8) Figure 9. Sample of the cartesian coordinates corresponding to detected photons (ɛ =.) over a restricted range to highlight the effect of Gaussian smearing, orginal poistions shown in (red) and smeared in (blue) Table. Sample of the mean x, y alterations < x >, < y > and thier respective ratios post Gaussian smearing for a range of decay numbers, N. N.B. Expections: [< y >/< x >] [y res/x res] ( ) N < x > ( m) < y > ( m) < y >/< x > in Table. confirms this tendency to the expected ratio as N.. Analysis An isotropic distribution requires the frequency of events to be proportional to solid angle Ω, the differential element of which is dω = sin(θ)dϕ = d(cosθ)dϕ, resulting in a uniformly spherical distribution. Ω frequencies are shown in Figure. and appear approximately uniform across all bins. In theory an infinite detector array should measure 5% of detections (discounting those with beyond the detector z > m) at a perfect quantum efficiency as statically half of the photons will propagate in directions with θ π/ for a sufficiently large number of decays ( O ). This is shown in Figure. as the detection fraction.5 as the detector becomes increasingly extensive in the x-y plane and ɛ. providing additional confirmation of isotropy. The distributions in Figure. resemble expectations and are consistent with Figures. and, depicticing increasingly extensive and isotropic photon detection as N increases. Symmetrical uniformity and isocontour definition (at base of each each D surface plot) illustrate the convergence toward continuity. The effect

5 Exercise : Random Numbers and Monte Carlo logn Ntotal = logn Ntotal = logn logn Ntotal = 5 Ntotal = logn Ntotal = logn Ntotal = logn logn Ntotal = 7 Ntotal = Figure. D (left) and D (right) spacial distributions of logarithmic photon detection frequency log (N ) across a m m detector array consisting of a square pixels (each with =.7, xres = cm and yres = cm) for a range of total number of decays Ntot = {,5,6,7}. Isocontours depict central elliptical aberation of the form (x/a) + (y/b) =, with eccentricity of e /.

6 6 J. F. Ider Chitham N=x 5 N=x 5 N=x 5 N=x 5 N=5x 5 Frequency π π π π Ω (sr) Figure. Histograms of solid angle Ω = πsin (θ/), illustrating an approximate increasingly isotropic distribution for N = {,,,, 5} 5. Detection Fraction N decays = Detector Size (N pixels ) ε Figure. Detection fraction as a function of detector size for a range of quantum efficiencies of the form a(n pixels b) c + d. of Gaussian smearing can be identified via the presence of an central elliptical aberration confirming the preliminary predictions of an asymmetry about reflections in x = y (.). ACKNOWLEDGEMENTS I thank my collaborators for assistance and C. Lucas for guidance throughout the duration of the course. This work made extensive use of GNU s GSL library. REFERENCES () Press, W.H., et al, Numerical Recipes in C, rd edn. Cambridge Univ. Press, Cambridge, 7 () Grimmett, G., Stirzaker, D., Probability & Random Processes, OUP, 99 () Devroye, L., Non-uniform random variate generation, st edn. Springer, 986 This paper has been typeset from a TEX/ L A TEX file prepared by the author.

Exercise 2: Partial Differential Equations

Exercise 2: Partial Differential Equations Exercise : Partial Differential Equations J F Ider Chitham Astrophysics Group, HH Wills Physics Laboratory, University of Bristol, Tyndall Avenue, Bristol BS8 TL Accepted March Received March ; in original

More information

Supplementary Information

Supplementary Information 1 Supplementary Information 3 Supplementary Figures 4 5 6 7 8 9 10 11 Supplementary Figure 1. Absorbing material placed between two dielectric media The incident electromagnetic wave propagates in stratified

More information

Poisson Variables. Robert DeSerio

Poisson Variables. Robert DeSerio Poisson Variables Robert DeSerio Poisson processes A Poisson process is one in which events are randomly distributed in time, space or some other variable with the number of events in any non-overlapping

More information

Absolute photon energy calibration using π 0 s

Absolute photon energy calibration using π 0 s Absolute photon energy calibration using π s A.D. Bukin BINP, Novosibirsk June 1, 25 BaBar Note #571 BaBar Note #582 analytically First idea of deconvolution algorithm Idea to use special fitting functions

More information

Monte Carlo Radiation Transfer I

Monte Carlo Radiation Transfer I Monte Carlo Radiation Transfer I Monte Carlo Photons and interactions Sampling from probability distributions Optical depths, isotropic emission, scattering Monte Carlo Basics Emit energy packet, hereafter

More information

Supplementary Figures

Supplementary Figures Supplementary Figures Supplementary Figure. X-ray diffraction pattern of CH 3 NH 3 PbI 3 film. Strong reflections of the () family of planes is characteristics of the preferred orientation of the perovskite

More information

The H.E.S.S. Standard Analysis Technique

The H.E.S.S. Standard Analysis Technique The H.E.S.S. Standard Analysis Technique Wystan Benbow for the H.E.S.S. Collaboration Max Planck Institut für Kernphysik Postfach 103980 D-69029 Heidelberg, Germany The High Energy Stereoscopic System

More information

Lecture 15. Theory of random processes Part III: Poisson random processes. Harrison H. Barrett University of Arizona

Lecture 15. Theory of random processes Part III: Poisson random processes. Harrison H. Barrett University of Arizona Lecture 15 Theory of random processes Part III: Poisson random processes Harrison H. Barrett University of Arizona 1 OUTLINE Poisson and independence Poisson and rarity; binomial selection Poisson point

More information

MASSACHUSETTS INSTITUTE OF TECHNOLOGY PHYSICS DEPARTMENT

MASSACHUSETTS INSTITUTE OF TECHNOLOGY PHYSICS DEPARTMENT G. Clark 7oct96 1 MASSACHUSETTS INSTITUTE OF TECHNOLOGY PHYSICS DEPARTMENT 8.13/8.14 Junior Laboratory STATISTICS AND ERROR ESTIMATION The purpose of this note is to explain the application of statistics

More information

Ø Set of mutually exclusive categories. Ø Classify or categorize subject. Ø No meaningful order to categorization.

Ø Set of mutually exclusive categories. Ø Classify or categorize subject. Ø No meaningful order to categorization. Statistical Tools in Evaluation HPS 41 Dr. Joe G. Schmalfeldt Types of Scores Continuous Scores scores with a potentially infinite number of values. Discrete Scores scores limited to a specific number

More information

Data Analysis I. Dr Martin Hendry, Dept of Physics and Astronomy University of Glasgow, UK. 10 lectures, beginning October 2006

Data Analysis I. Dr Martin Hendry, Dept of Physics and Astronomy University of Glasgow, UK. 10 lectures, beginning October 2006 Astronomical p( y x, I) p( x, I) p ( x y, I) = p( y, I) Data Analysis I Dr Martin Hendry, Dept of Physics and Astronomy University of Glasgow, UK 10 lectures, beginning October 2006 4. Monte Carlo Methods

More information

1. Kinematics, cross-sections etc

1. Kinematics, cross-sections etc 1. Kinematics, cross-sections etc A study of kinematics is of great importance to any experiment on particle scattering. It is necessary to interpret your measurements, but at an earlier stage to determine

More information

Methods of Data Analysis Random numbers, Monte Carlo integration, and Stochastic Simulation Algorithm (SSA / Gillespie)

Methods of Data Analysis Random numbers, Monte Carlo integration, and Stochastic Simulation Algorithm (SSA / Gillespie) Methods of Data Analysis Random numbers, Monte Carlo integration, and Stochastic Simulation Algorithm (SSA / Gillespie) Week 1 1 Motivation Random numbers (RNs) are of course only pseudo-random when generated

More information

Lecture 5: Random numbers and Monte Carlo (Numerical Recipes, Chapter 7) Motivations for generating random numbers

Lecture 5: Random numbers and Monte Carlo (Numerical Recipes, Chapter 7) Motivations for generating random numbers Lecture 5: Random numbers and Monte Carlo (Numerical Recipes, Chapter 7) Motivations for generating random numbers To sample a function in a statistically controlled manner (i.e. for Monte Carlo integration)

More information

Physics 403 Probability Distributions II: More Properties of PDFs and PMFs

Physics 403 Probability Distributions II: More Properties of PDFs and PMFs Physics 403 Probability Distributions II: More Properties of PDFs and PMFs Segev BenZvi Department of Physics and Astronomy University of Rochester Table of Contents 1 Last Time: Common Probability Distributions

More information

Brandon C. Kelly (Harvard Smithsonian Center for Astrophysics)

Brandon C. Kelly (Harvard Smithsonian Center for Astrophysics) Brandon C. Kelly (Harvard Smithsonian Center for Astrophysics) Probability quantifies randomness and uncertainty How do I estimate the normalization and logarithmic slope of a X ray continuum, assuming

More information

AT2 Neuromodeling: Problem set #3 SPIKE TRAINS

AT2 Neuromodeling: Problem set #3 SPIKE TRAINS AT2 Neuromodeling: Problem set #3 SPIKE TRAINS Younesse Kaddar PROBLEM 1: Poisson spike trains Link of the ipython notebook for the code Brain neuron emit spikes seemingly randomly: we will aim to model

More information

Polarization Correlation in the Gamma- Gamma Decay of Positronium

Polarization Correlation in the Gamma- Gamma Decay of Positronium Polarization Correlation in the Gamma- Gamma Decay of Positronium Bin LI Department of Physics & Astronomy, University of Pittsburgh, PA 56, U.S.A April 5, Introduction Positronium is an unstable bound

More information

Electrodynamics of Radiation Processes

Electrodynamics of Radiation Processes Electrodynamics of Radiation Processes 7. Emission from relativistic particles (contd) & Bremsstrahlung http://www.astro.rug.nl/~etolstoy/radproc/ Chapter 4: Rybicki&Lightman Sections 4.8, 4.9 Chapter

More information

Simulation. Where real stuff starts

Simulation. Where real stuff starts 1 Simulation Where real stuff starts ToC 1. What is a simulation? 2. Accuracy of output 3. Random Number Generators 4. How to sample 5. Monte Carlo 6. Bootstrap 2 1. What is a simulation? 3 What is a simulation?

More information

Variational Principal Components

Variational Principal Components Variational Principal Components Christopher M. Bishop Microsoft Research 7 J. J. Thomson Avenue, Cambridge, CB3 0FB, U.K. cmbishop@microsoft.com http://research.microsoft.com/ cmbishop In Proceedings

More information

Sample Spectroscopy System Hardware

Sample Spectroscopy System Hardware Semiconductor Detectors vs. Scintillator+PMT Detectors Semiconductors are emerging technology - Scint.PMT systems relatively unchanged in 50 years. NaI(Tl) excellent for single-photon, new scintillation

More information

A MONTE CARLO SIMULATION OF COMPTON SUPPRESSION FOR NEUTRON ACTIVATION ANALYSIS. Joshua Frye Adviser Chris Grant 8/24/2012 ABSTRACT

A MONTE CARLO SIMULATION OF COMPTON SUPPRESSION FOR NEUTRON ACTIVATION ANALYSIS. Joshua Frye Adviser Chris Grant 8/24/2012 ABSTRACT A MONTE CARLO SIMULATION OF COMPTON SUPPRESSION FOR NEUTRON ACTIVATION ANALYSIS Joshua Frye Adviser Chris Grant 8/24/2012 ABSTRACT A Monte Carlo simulation has been developed using the Geant4 software

More information

Error analysis for efficiency

Error analysis for efficiency Glen Cowan RHUL Physics 28 July, 2008 Error analysis for efficiency To estimate a selection efficiency using Monte Carlo one typically takes the number of events selected m divided by the number generated

More information

Applied Nuclear Physics (Fall 2006) Lecture 19 (11/22/06) Gamma Interactions: Compton Scattering

Applied Nuclear Physics (Fall 2006) Lecture 19 (11/22/06) Gamma Interactions: Compton Scattering .101 Applied Nuclear Physics (Fall 006) Lecture 19 (11//06) Gamma Interactions: Compton Scattering References: R. D. Evans, Atomic Nucleus (McGraw-Hill New York, 1955), Chaps 3 5.. W. E. Meyerhof, Elements

More information

Comparison of Modern Stochastic Optimization Algorithms

Comparison of Modern Stochastic Optimization Algorithms Comparison of Modern Stochastic Optimization Algorithms George Papamakarios December 214 Abstract Gradient-based optimization methods are popular in machine learning applications. In large-scale problems,

More information

Recall the Basics of Hypothesis Testing

Recall the Basics of Hypothesis Testing Recall the Basics of Hypothesis Testing The level of significance α, (size of test) is defined as the probability of X falling in w (rejecting H 0 ) when H 0 is true: P(X w H 0 ) = α. H 0 TRUE H 1 TRUE

More information

Bayesian search for other Earths

Bayesian search for other Earths Bayesian search for other Earths Low-mass planets orbiting nearby M dwarfs Mikko Tuomi University of Hertfordshire, Centre for Astrophysics Research Email: mikko.tuomi@utu.fi Presentation, 19.4.2013 1

More information

GRADUATE WRITTEN EXAMINATION. Fall 2018 PART I

GRADUATE WRITTEN EXAMINATION. Fall 2018 PART I University of Minnesota School of Physics and Astronomy GRADUATE WRITTEN EXAMINATION Fall 2018 PART I Monday, August 20 th, 2018 9:00 am to 1:00 pm Part 1 of this exam consists of 10 problems of equal

More information

Solutions for Exercise session I

Solutions for Exercise session I Solutions for Exercise session I 1. The maximally polarisation-entangled photon state can be written as Ψ = 1 ( H 1 V V 1 H ). Show that the state is invariant (i.e. still maximally entangled) after a

More information

6. MC data analysis. Doing experiments or complex simulations can be quite time-consuming, expensive, or both.

6. MC data analysis. Doing experiments or complex simulations can be quite time-consuming, expensive, or both. 6. MC data analysis [Numerical Recipes 15.6, own derivation] Monte Carlo simulations can be a highly useful tool in the analysis of experimental data, or data produced by other computer simulations. Doing

More information

Statistics, Data Analysis, and Simulation SS 2017

Statistics, Data Analysis, and Simulation SS 2017 Statistics, Data Analysis, and Simulation SS 2017 08.128.730 Statistik, Datenanalyse und Simulation Dr. Michael O. Distler Mainz, 27. April 2017 Dr. Michael O. Distler

More information

Review Notes for IB Standard Level Math

Review Notes for IB Standard Level Math Review Notes for IB Standard Level Math 1 Contents 1 Algebra 8 1.1 Rules of Basic Operations............................... 8 1.2 Rules of Roots..................................... 8 1.3 Rules of Exponents...................................

More information

Chapter 2. Random Variable. Define single random variables in terms of their PDF and CDF, and calculate moments such as the mean and variance.

Chapter 2. Random Variable. Define single random variables in terms of their PDF and CDF, and calculate moments such as the mean and variance. Chapter 2 Random Variable CLO2 Define single random variables in terms of their PDF and CDF, and calculate moments such as the mean and variance. 1 1. Introduction In Chapter 1, we introduced the concept

More information

Multiple Random Variables

Multiple Random Variables Multiple Random Variables Joint Probability Density Let X and Y be two random variables. Their joint distribution function is F ( XY x, y) P X x Y y. F XY ( ) 1, < x

More information

6. MC data analysis. [Numerical Recipes 15.6, own derivation] Basics of Monte Carlo simulations, Kai Nordlund

6. MC data analysis. [Numerical Recipes 15.6, own derivation] Basics of Monte Carlo simulations, Kai Nordlund 6. MC data analysis [Numerical Recipes 15.6, own derivation] Monte Carlo simulations can be a highly useful tool in the analysis of experimental data, or data produced by other computer simulations. The

More information

Statistical Methods for Astronomy

Statistical Methods for Astronomy Statistical Methods for Astronomy If your experiment needs statistics, you ought to have done a better experiment. -Ernest Rutherford Lecture 1 Lecture 2 Why do we need statistics? Definitions Statistical

More information

Statistics for Data Analysis. Niklaus Berger. PSI Practical Course Physics Institute, University of Heidelberg

Statistics for Data Analysis. Niklaus Berger. PSI Practical Course Physics Institute, University of Heidelberg Statistics for Data Analysis PSI Practical Course 2014 Niklaus Berger Physics Institute, University of Heidelberg Overview You are going to perform a data analysis: Compare measured distributions to theoretical

More information

Monte Carlo radiation transport codes

Monte Carlo radiation transport codes Monte Carlo radiation transport codes How do they work? Michel Maire (Lapp/Annecy) 23/05/2007 introduction to Monte Carlo radiation transport codes 1 Decay in flight (1) An unstable particle have a time

More information

1 Statistics Aneta Siemiginowska a chapter for X-ray Astronomy Handbook October 2008

1 Statistics Aneta Siemiginowska a chapter for X-ray Astronomy Handbook October 2008 1 Statistics Aneta Siemiginowska a chapter for X-ray Astronomy Handbook October 2008 1.1 Introduction Why do we need statistic? Wall and Jenkins (2003) give a good description of the scientific analysis

More information

Detection of X-Rays. Solid state detectors Proportional counters Microcalorimeters Detector characteristics

Detection of X-Rays. Solid state detectors Proportional counters Microcalorimeters Detector characteristics Detection of X-Rays Solid state detectors Proportional counters Microcalorimeters Detector characteristics Solid State X-ray Detectors X-ray interacts in material to produce photoelectrons which are collected

More information

Fall 2012 Analysis of Experimental Measurements B. Eisenstein/rev. S. Errede

Fall 2012 Analysis of Experimental Measurements B. Eisenstein/rev. S. Errede Hypothesis Testing: Suppose we have two or (in general) more simple hypotheses which can describe a set of data Simple means explicitly defined, so if parameters have to be fitted, that has already been

More information

Detectors in Nuclear Physics: Monte Carlo Methods. Dr. Andrea Mairani. Lectures I-II

Detectors in Nuclear Physics: Monte Carlo Methods. Dr. Andrea Mairani. Lectures I-II Detectors in Nuclear Physics: Monte Carlo Methods Dr. Andrea Mairani Lectures I-II INTRODUCTION Sampling from a probability distribution Sampling from a probability distribution X λ Sampling from a probability

More information

Modern Methods of Data Analysis - WS 07/08

Modern Methods of Data Analysis - WS 07/08 Modern Methods of Data Analysis Lecture V (12.11.07) Contents: Central Limit Theorem Uncertainties: concepts, propagation and properties Central Limit Theorem Consider the sum X of n independent variables,

More information

Introduction to Probability

Introduction to Probability LECTURE NOTES Course 6.041-6.431 M.I.T. FALL 2000 Introduction to Probability Dimitri P. Bertsekas and John N. Tsitsiklis Professors of Electrical Engineering and Computer Science Massachusetts Institute

More information

Ø Set of mutually exclusive categories. Ø Classify or categorize subject. Ø No meaningful order to categorization.

Ø Set of mutually exclusive categories. Ø Classify or categorize subject. Ø No meaningful order to categorization. Statistical Tools in Evaluation HPS 41 Fall 213 Dr. Joe G. Schmalfeldt Types of Scores Continuous Scores scores with a potentially infinite number of values. Discrete Scores scores limited to a specific

More information

6. LIGHT SCATTERING 6.1 The first Born approximation

6. LIGHT SCATTERING 6.1 The first Born approximation 6. LIGHT SCATTERING 6.1 The first Born approximation In many situations, light interacts with inhomogeneous systems, in which case the generic light-matter interaction process is referred to as scattering

More information

Photoneutron reactions studies at ELI-NP using a direct neutron multiplicity sorting method Dan Filipescu

Photoneutron reactions studies at ELI-NP using a direct neutron multiplicity sorting method Dan Filipescu EUROPEAN UNION GOVERNMENT OF ROMANIA Sectoral Operational Programme Increase of Economic Competitiveness Investments for Your Future Structural Instruments 2007-2013 Extreme Light Infrastructure Nuclear

More information

EE/CpE 345. Modeling and Simulation. Fall Class 9

EE/CpE 345. Modeling and Simulation. Fall Class 9 EE/CpE 345 Modeling and Simulation Class 9 208 Input Modeling Inputs(t) Actual System Outputs(t) Parameters? Simulated System Outputs(t) The input data is the driving force for the simulation - the behavior

More information

Statistics for Data Analysis a toolkit for the (astro)physicist

Statistics for Data Analysis a toolkit for the (astro)physicist Statistics for Data Analysis a toolkit for the (astro)physicist Likelihood in Astrophysics Denis Bastieri Padova, February 21 st 2018 RECAP The main product of statistical inference is the pdf of the model

More information

Monte Carlo and cold gases. Lode Pollet.

Monte Carlo and cold gases. Lode Pollet. Monte Carlo and cold gases Lode Pollet lpollet@physics.harvard.edu 1 Outline Classical Monte Carlo The Monte Carlo trick Markov chains Metropolis algorithm Ising model critical slowing down Quantum Monte

More information

PMT Charge Response and Reconstruction of Position and Energy for a Reactor θ 13 Experiment

PMT Charge Response and Reconstruction of Position and Energy for a Reactor θ 13 Experiment PMT Charge Response and Reconstruction of Position and Energy for a Reactor θ 13 Experiment Josh Klein April 11, 2004 1 Introduction I have taken a first look at the effects of the PMT charge response

More information

MTH739U/P: Topics in Scientific Computing Autumn 2016 Week 6

MTH739U/P: Topics in Scientific Computing Autumn 2016 Week 6 MTH739U/P: Topics in Scientific Computing Autumn 16 Week 6 4.5 Generic algorithms for non-uniform variates We have seen that sampling from a uniform distribution in [, 1] is a relatively straightforward

More information

The Compton Effect. Martha Buckley MIT Department of Physics, Cambridge, MA (Dated: November 26, 2002)

The Compton Effect. Martha Buckley MIT Department of Physics, Cambridge, MA (Dated: November 26, 2002) The Compton Effect Martha Buckley MIT Department of Physics, Cambridge, MA 02139 marthab@mit.edu (Dated: November 26, 2002) We measured the angular dependence of the energies of 661.6 kev photons scattered

More information

Statistics, Data Analysis, and Simulation SS 2013

Statistics, Data Analysis, and Simulation SS 2013 Statistics, Data Analysis, and Simulation SS 213 8.128.73 Statistik, Datenanalyse und Simulation Dr. Michael O. Distler Mainz, 23. April 213 What we ve learned so far Fundamental

More information

Lecture 13 and 14: Bayesian estimation theory

Lecture 13 and 14: Bayesian estimation theory 1 Lecture 13 and 14: Bayesian estimation theory Spring 2012 - EE 194 Networked estimation and control (Prof. Khan) March 26 2012 I. BAYESIAN ESTIMATORS Mother Nature conducts a random experiment that generates

More information

3 a = 3 b c 2 = a 2 + b 2 = 2 2 = 4 c 2 = 3b 2 + b 2 = 4b 2 = 4 b 2 = 1 b = 1 a = 3b = 3. x 2 3 y2 1 = 1.

3 a = 3 b c 2 = a 2 + b 2 = 2 2 = 4 c 2 = 3b 2 + b 2 = 4b 2 = 4 b 2 = 1 b = 1 a = 3b = 3. x 2 3 y2 1 = 1. MATH 222 LEC SECOND MIDTERM EXAM THU NOV 8 PROBLEM ( 5 points ) Find the standard-form equation for the hyperbola which has its foci at F ± (±2, ) and whose asymptotes are y ± 3 x The calculations b a

More information

Simulation. Where real stuff starts

Simulation. Where real stuff starts Simulation Where real stuff starts March 2019 1 ToC 1. What is a simulation? 2. Accuracy of output 3. Random Number Generators 4. How to sample 5. Monte Carlo 6. Bootstrap 2 1. What is a simulation? 3

More information

PH2200 Practice Final Exam Summer 2003

PH2200 Practice Final Exam Summer 2003 INSTRUCTIONS 1. Write your name and student identification number on the answer sheet. 2. Please cover your answer sheet at all times. 3. This is a closed book exam. You may use the PH2200 formula sheet

More information

Cosmology & CMB. Set5: Data Analysis. Davide Maino

Cosmology & CMB. Set5: Data Analysis. Davide Maino Cosmology & CMB Set5: Data Analysis Davide Maino Gaussian Statistics Statistical isotropy states only two-point correlation function is needed and it is related to power spectrum Θ(ˆn) = lm Θ lm Y lm (ˆn)

More information

Investigation of Possible Biases in Tau Neutrino Mass Limits

Investigation of Possible Biases in Tau Neutrino Mass Limits Investigation of Possible Biases in Tau Neutrino Mass Limits Kyle Armour Departments of Physics and Mathematics, University of California, San Diego, La Jolla, CA 92093 (Dated: August 8, 2003) We study

More information

A Study of Covariances within Basic and Extended Kalman Filters

A Study of Covariances within Basic and Extended Kalman Filters A Study of Covariances within Basic and Extended Kalman Filters David Wheeler Kyle Ingersoll December 2, 2013 Abstract This paper explores the role of covariance in the context of Kalman filters. The underlying

More information

Non-linear least-squares and chemical kinetics. An improved method to analyse monomer-excimer decay data

Non-linear least-squares and chemical kinetics. An improved method to analyse monomer-excimer decay data Journal of Mathematical Chemistry 21 (1997) 131 139 131 Non-linear least-squares and chemical kinetics. An improved method to analyse monomer-excimer decay data J.P.S. Farinha a, J.M.G. Martinho a and

More information

PHYSICS WORKSHEET CLASS : XII

PHYSICS WORKSHEET CLASS : XII PHYSICS WORKSHEET CLASS : XII Chapter 2 Topic: Electrostatic potential and Capacitance LEVEL 1 1. Define electric potential. Derive the expression for the electric potential at any point due to a point

More information

arxiv: v1 [astro-ph.he] 7 Mar 2018

arxiv: v1 [astro-ph.he] 7 Mar 2018 Extracting a less model dependent cosmic ray composition from X max distributions Simon Blaess, Jose A. Bellido, and Bruce R. Dawson Department of Physics, University of Adelaide, Adelaide, Australia arxiv:83.v

More information

DVCS with CLAS. Elton S. Smith. Jefferson Lab. Conference on Intersections of Particle and Nuclear Physics New York, Elton S.

DVCS with CLAS. Elton S. Smith. Jefferson Lab. Conference on Intersections of Particle and Nuclear Physics New York, Elton S. DVCS with CLAS Elton S. Smith Jefferson Lab Conference on Intersections of Particle and Nuclear Physics New York, 2003 Elton S. Smith 1 Deeply Virtual Compton Scattering Inclusive Scattering Forward Compton

More information

functions Poisson distribution Normal distribution Arbitrary functions

functions Poisson distribution Normal distribution Arbitrary functions Physics 433: Computational Physics Lecture 6 Random number distributions Generation of random numbers of various distribuition functions Normal distribution Poisson distribution Arbitrary functions Random

More information

If we want to analyze experimental or simulated data we might encounter the following tasks:

If we want to analyze experimental or simulated data we might encounter the following tasks: Chapter 1 Introduction If we want to analyze experimental or simulated data we might encounter the following tasks: Characterization of the source of the signal and diagnosis Studying dependencies Prediction

More information

Scientific Computing: Monte Carlo

Scientific Computing: Monte Carlo Scientific Computing: Monte Carlo Aleksandar Donev Courant Institute, NYU 1 donev@courant.nyu.edu 1 Course MATH-GA.2043 or CSCI-GA.2112, Spring 2012 April 5th and 12th, 2012 A. Donev (Courant Institute)

More information

Statistics, Data Analysis, and Simulation SS 2015

Statistics, Data Analysis, and Simulation SS 2015 Statistics, Data Analysis, and Simulation SS 2015 08.128.730 Statistik, Datenanalyse und Simulation Dr. Michael O. Distler Mainz, 27. April 2015 Dr. Michael O. Distler

More information

Measuring Cosmic Ray Muon Decay Constant and Flux

Measuring Cosmic Ray Muon Decay Constant and Flux WJP, PHY381 (2015) Wabash Journal of Physics v3.3, p.1 Measuring Cosmic Ray Muon Decay Constant and Flux R.C. Dennis, D.T. Tran, J. Qi, and J. Brown Department of Physics, Wabash College, Crawfordsville,

More information

Calibration of the AGILE Gamma Ray Imaging Detector

Calibration of the AGILE Gamma Ray Imaging Detector Calibration of the AGILE Gamma Ray Imaging Detector Andrew Chen on behalf of the AGILE Team April 11, 2011 AGILE Astrorivelatore Gamma ad Immagini LEggero Italian Space Agency (ASI) small mission Participation

More information

Lecture 3. G. Cowan. Lecture 3 page 1. Lectures on Statistical Data Analysis

Lecture 3. G. Cowan. Lecture 3 page 1. Lectures on Statistical Data Analysis Lecture 3 1 Probability (90 min.) Definition, Bayes theorem, probability densities and their properties, catalogue of pdfs, Monte Carlo 2 Statistical tests (90 min.) general concepts, test statistics,

More information

Problem Set Number 02, j/2.036j MIT (Fall 2018)

Problem Set Number 02, j/2.036j MIT (Fall 2018) Problem Set Number 0, 18.385j/.036j MIT (Fall 018) Rodolfo R. Rosales (MIT, Math. Dept., room -337, Cambridge, MA 0139) September 6, 018 Due October 4, 018. Turn it in (by 3PM) at the Math. Problem Set

More information

Counting Photons to Calibrate a Photometer for Stellar Intensity Interferometry

Counting Photons to Calibrate a Photometer for Stellar Intensity Interferometry Counting Photons to Calibrate a Photometer for Stellar Intensity Interferometry A Senior Project Presented to the Department of Physics California Polytechnic State University, San Luis Obispo In Partial

More information

6. Advanced Numerical Methods. Monte Carlo Methods

6. Advanced Numerical Methods. Monte Carlo Methods 6. Advanced Numerical Methods Part 1: Part : Monte Carlo Methods Fourier Methods Part 1: Monte Carlo Methods 1. Uniform random numbers Generating uniform random numbers, drawn from the pdf U[0,1], is fairly

More information

- 1 - θ 1. n 1. θ 2. mirror. object. image

- 1 - θ 1. n 1. θ 2. mirror. object. image TEST 5 (PHY 50) 1. a) How will the ray indicated in the figure on the following page be reflected by the mirror? (Be accurate!) b) Explain the symbols in the thin lens equation. c) Recall the laws governing

More information

Geosynchrotron radio emission from CORSIKAsimulated

Geosynchrotron radio emission from CORSIKAsimulated Journal of Physics: Conference Series Geosynchrotron radio emission from CORSIKAsimulated air showers To cite this article: T Huege et al 27 J. Phys.: Conf. Ser. 81 122 Related content - Radio emission

More information

EE/CpE 345. Modeling and Simulation. Fall Class 10 November 18, 2002

EE/CpE 345. Modeling and Simulation. Fall Class 10 November 18, 2002 EE/CpE 345 Modeling and Simulation Class 0 November 8, 2002 Input Modeling Inputs(t) Actual System Outputs(t) Parameters? Simulated System Outputs(t) The input data is the driving force for the simulation

More information

Compute the Fourier transform on the first register to get x {0,1} n x 0.

Compute the Fourier transform on the first register to get x {0,1} n x 0. CS 94 Recursive Fourier Sampling, Simon s Algorithm /5/009 Spring 009 Lecture 3 1 Review Recall that we can write any classical circuit x f(x) as a reversible circuit R f. We can view R f as a unitary

More information

Proximity Decay and Tidal Effects

Proximity Decay and Tidal Effects Proximity Decay and Tidal Effects A. B. McIntosh,S. Hudan, C.J. Metelko, N. Peters, J. Black, RdS Dept of Chemistry and IUCF, Indiana University July 16 22 1994: http://www2.jpl.nasa.gov/sl9/ Comet P/Shoemaker-Levy

More information

Statistical Methods for Particle Physics Lecture 1: parameter estimation, statistical tests

Statistical Methods for Particle Physics Lecture 1: parameter estimation, statistical tests Statistical Methods for Particle Physics Lecture 1: parameter estimation, statistical tests http://benasque.org/2018tae/cgi-bin/talks/allprint.pl TAE 2018 Benasque, Spain 3-15 Sept 2018 Glen Cowan Physics

More information

A6523 Signal Modeling, Statistical Inference and Data Mining in Astrophysics Spring 2011

A6523 Signal Modeling, Statistical Inference and Data Mining in Astrophysics Spring 2011 A6523 Signal Modeling, Statistical Inference and Data Mining in Astrophysics Spring 2011 Reading Chapter 5 (continued) Lecture 8 Key points in probability CLT CLT examples Prior vs Likelihood Box & Tiao

More information

Statistical Methods in Particle Physics

Statistical Methods in Particle Physics Statistical Methods in Particle Physics Lecture 3 October 29, 2012 Silvia Masciocchi, GSI Darmstadt s.masciocchi@gsi.de Winter Semester 2012 / 13 Outline Reminder: Probability density function Cumulative

More information

Chapter 4. Repeated Trials. 4.1 Introduction. 4.2 Bernoulli Trials

Chapter 4. Repeated Trials. 4.1 Introduction. 4.2 Bernoulli Trials Chapter 4 Repeated Trials 4.1 Introduction Repeated indepentent trials in which there can be only two outcomes are called Bernoulli trials in honor of James Bernoulli (1654-1705). As we shall see, Bernoulli

More information

Monte Carlo Simulations

Monte Carlo Simulations Monte Carlo Simulations What are Monte Carlo Simulations and why ones them? Pseudo Random Number generators Creating a realization of a general PDF The Bootstrap approach A real life example: LOFAR simulations

More information

1 Fundamentals. 1.1 Overview. 1.2 Units: Physics 704 Spring 2018

1 Fundamentals. 1.1 Overview. 1.2 Units: Physics 704 Spring 2018 Physics 704 Spring 2018 1 Fundamentals 1.1 Overview The objective of this course is: to determine and fields in various physical systems and the forces and/or torques resulting from them. The domain of

More information

Physics 6720 Introduction to Statistics April 4, 2017

Physics 6720 Introduction to Statistics April 4, 2017 Physics 6720 Introduction to Statistics April 4, 2017 1 Statistics of Counting Often an experiment yields a result that can be classified according to a set of discrete events, giving rise to an integer

More information

ANTENNAS. Vector and Scalar Potentials. Maxwell's Equations. E = jωb. H = J + jωd. D = ρ (M3) B = 0 (M4) D = εe

ANTENNAS. Vector and Scalar Potentials. Maxwell's Equations. E = jωb. H = J + jωd. D = ρ (M3) B = 0 (M4) D = εe ANTENNAS Vector and Scalar Potentials Maxwell's Equations E = jωb H = J + jωd D = ρ B = (M) (M) (M3) (M4) D = εe B= µh For a linear, homogeneous, isotropic medium µ and ε are contant. Since B =, there

More information

8 STOCHASTIC SIMULATION

8 STOCHASTIC SIMULATION 8 STOCHASTIC SIMULATIO 59 8 STOCHASTIC SIMULATIO Whereas in optimization we seek a set of parameters x to minimize a cost, or to maximize a reward function J( x), here we pose a related but different question.

More information

Department of Electrical- and Information Technology. ETS061 Lecture 3, Verification, Validation and Input

Department of Electrical- and Information Technology. ETS061 Lecture 3, Verification, Validation and Input ETS061 Lecture 3, Verification, Validation and Input Verification and Validation Real system Validation Model Verification Measurements Verification Break model down into smaller bits and test each bit

More information

Bias-Variance Error Bounds for Temporal Difference Updates

Bias-Variance Error Bounds for Temporal Difference Updates Bias-Variance Bounds for Temporal Difference Updates Michael Kearns AT&T Labs mkearns@research.att.com Satinder Singh AT&T Labs baveja@research.att.com Abstract We give the first rigorous upper bounds

More information

Electromagnetic Radiation. Physical Principles of Remote Sensing

Electromagnetic Radiation. Physical Principles of Remote Sensing Electromagnetic Radiation Physical Principles of Remote Sensing Outline for 4/3/2003 Properties of electromagnetic radiation The electromagnetic spectrum Spectral emissivity Radiant temperature vs. kinematic

More information

MCRT: L4 A Monte Carlo Scattering Code

MCRT: L4 A Monte Carlo Scattering Code MCRT: L4 A Monte Carlo Scattering Code Plane parallel scattering slab Optical depths & physical distances Emergent flux & intensity Internal intensity moments Constant density slab, vertical optical depth

More information

Efficient sorting of orbital angular momentum states of light

Efficient sorting of orbital angular momentum states of light CHAPTER 6 Efficient sorting of orbital angular momentum states of light We present a method to efficiently sort orbital angular momentum (OAM) states of light using two static optical elements. The optical

More information

Adjustments for the Display of Quantized Ion Channel Dwell Times in Histograms with Logarithmic Bins

Adjustments for the Display of Quantized Ion Channel Dwell Times in Histograms with Logarithmic Bins 662 Biophysical Journal Volume 78 February 2000 662 667 Adjustments for the Display of Quantized Ion Channel Dwell Times in Histograms with Logarithmic Bins J. Alex Stark and Stephen B. Hladky National

More information

Bethe-Block. Stopping power of positive muons in copper vs βγ = p/mc. The slight dependence on M at highest energies through T max

Bethe-Block. Stopping power of positive muons in copper vs βγ = p/mc. The slight dependence on M at highest energies through T max Bethe-Block Stopping power of positive muons in copper vs βγ = p/mc. The slight dependence on M at highest energies through T max can be used for PID but typically de/dx depend only on β (given a particle

More information

A Review of Basic Monte Carlo Methods

A Review of Basic Monte Carlo Methods A Review of Basic Monte Carlo Methods Julian Haft May 9, 2014 Introduction One of the most powerful techniques in statistical analysis developed in this past century is undoubtedly that of Monte Carlo

More information

CsI Calorimeter for KOTO experiment

CsI Calorimeter for KOTO experiment PROCEEDINGSof CHEF23 CsI Calorimeter for KOTO experiment Department of Physics, Osaka University E-mail: sato@champ.hep.sci.osaka-u.ac.jp The J-PARC KOTO experiment searches for K L π ν ν decay by observing

More information

Supplementary Information: Lifetime measurements well below the diffraction limit

Supplementary Information: Lifetime measurements well below the diffraction limit Supplementary Information: Lifetime measurements well below the diffraction limit S. Meuret, L. H. G. Tizei, T.Auzelle, B. Daudin, B. Gayral and M. Kociak 1 High Angle Annular Dark Field acquired during

More information