DOCSLIB.ORG

Lecture 20 Randomness and Monte Carlo

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Lecture 20 Randomness and Monte Carlo Lecture 20 Randomness and Monte Carlo J. Chaudhry Department of Mathematics and Statistics University of New Mexico J. Chaudhry (UNM) CS 357 1 / 41 What we’ll do: Random number generators Monte-Carlo integration Monte-Carlo simulation Our testing and our methods often rely on “randomness” and the impact of finite precision Random input Random perturbations Random sequences J. Chaudhry (UNM) CS 357 2 / 41 Randomness Randomness ≈ unpredictability One view: a sequence is random if it has no shorter description Physical processes, such as flipping a coin or tossing dice, are deterministic with enough information about the governing equations and initial conditions. But even for deterministic systems, sensitivity to the initial conditions can render the behavior practically unpredictable. we need random simulation methods J. Chaudhry (UNM) CS 357 3 / 41 Quasi-Random Sequences For some applications, reasonable uniform coverage of the sample is more important than the “randomness” True random samples often exhibit clumping Perfectly uniform samples uses a uniform grid, but does not scale well at high dimensions quasi-random sequences attempt randomness while maintaining coverage J. Chaudhry (UNM) CS 357 4 / 41 Quasi-Random Sequences quasi random sequences are not random, but give random appearance by design, the points avoid each other, resulting in no clumping http://dilbert.com/strips/comic/2001-10-25/ http://www.xkcd.com/221/ J. Chaudhry (UNM) CS 357 5 / 41 Randomness is easy, right? In May, 2008, Debian announced a vulnerability with OpenSSL: the OpenSSL pseudo-random number generator I the seeding process was compromised (2 years) I only 32,767 possible keys I seeding based on process ID (this is not entropy!) I all SSL and SSH keys from 9/2006 - 5/2008 regenerated I all certificates recertified Cryptographically secure pseudorandom number generator (CSPRNG) are necessary for some apps Other apps rely less on true randomness J. Chaudhry (UNM) CS 357 6 / 41 Repeatability With unpredictability, true randomness is not repeatable ...but lack of repeatability makes testing/debugging difficult So we want repeatability, but also independence of the trials 1 >> rand('seed',1234) 2 >> rand(10,1) J. Chaudhry (UNM) CS 357 7 / 41 Pseudorandom Numbers Computer algorithms for random number generations are deterministic ...but may have long periodicity (a long time until an apparent pattern emerges) These sequences are labeled pseudorandom Pseudorandom sequences are predictable and reproducible (this is mostly good) J. Chaudhry (UNM) CS 357 8 / 41 Random Number Generators Properties of a good random number generator: Random pattern: passes statistical tests of randomness Long period: long time before repeating Efficiency: executes rapidly and with low storage Repeatability: same sequence is generated using same initial states Portability: same sequences are generated on different architectures J. Chaudhry (UNM) CS 357 9 / 41 Random Number Generators Early attempts relied on complexity to ensure randomness “midsquare” method: square each member of a sequence and take the middle portion of the results as the next member of the sequence ...simple methods with a statistical basis are preferable J. Chaudhry (UNM) CS 357 10 / 41 Linear Congruential Generators (LCGs) Congruential random number generators are of the form: xk = (axk-1 + c)( mod m) where a and c are integers given as input. x0 is called the seed Integer m - 1 is the largest integer representable Quality depends on a and c. The period will be at most m. Example Let a = 13, c = 0, m = 31, and x0 = 1. 1, 13, 14, 27, 10, 6,... This is a permutation of integers from 1,..., 30, so the period is m - 1. J. Chaudhry (UNM) CS 357 11 / 41 History From C. Moler, NCM IBM used Scientific Subroutine Package (SSP) in the 1960’s the mainframes. Their random generator, rnd used a = 65539, c = 0, and m = 231. arithmetic mod 231 is done quickly with 32 bit words. multiplication can be done quickly with a = 216 + 3 with a shift and short add. Notice (mod m): xk+2 = 6xk+1 - 9xk ...strong correlation among three successive integers J. Chaudhry (UNM) CS 357 12 / 41 History From C. Moler, NCM Matlab used a = 75, c = 0, and m = 231 - 1 for a while period is m - 1. this is no longer sufficient J. Chaudhry (UNM) CS 357 13 / 41 Linear Congruential Generators (LCGs) – Summary sensitive to a and c be careful with supplied random functions on your system period is m standard division is necessary if generating floating points in [0, 1). J. Chaudhry (UNM) CS 357 14 / 41 what’s used? Two popular methods: 1. rand() in Unix uses LCG with a = 1103515245, c = 12345, m = 231. 2. Method of Marsaglia (period ≈ 21430). 1 Initialize x0,..., x3 and c to random values given a seed 2 3 Let s = 2111111111xn-4 + 1492xn-31776xn-2 + 5115xn-1 + c 4 32 5 Compute xn = s mod 2 6 32 7 c = floor(s=2 ) In general, the digits in random numbers are not themselves random...some patterns reoccur much more often. J. Chaudhry (UNM) CS 357 15 / 41 Fibonacci produce floating-point random numbers directly using differences, sums, or products. Typical subtractive generator: xk = xk-17 - xk-5 with “lags” of 17 and 5. Lags much be chosen very carefully negative results need fixing more storage needed than congruential generators no division needed very very good statistical properties long periods since repetition does not imply a period J. Chaudhry (UNM) CS 357 16 / 41 Discrete random variables random variable x values: x1,..., xn (the set of values is called sample space) n probabilities p1,..., pn with i=1 pi = 1 P throwing a die (1-based index) values: x1 = 1, x2 = 2,..., x6 = 6 probabilities pi = 1=6 J. Chaudhry (UNM) CS 357 17 / 41 Expected value and variance expected value: average value of the variable n E[x] = xjpj = Xj 1 variance: variation from the average σ2[x] = E[(x - E[x])2] = E[x2]- E[x]2 Standard deviation: square root of variance. denoted by σ[x] throwing a die expected value: E[x] = (1 + 2 + ··· + 6)=6 = 3.5 variance: σ2[x] = 2.916 J. Chaudhry (UNM) CS 357 18 / 41 Estimated E[x] to estimate the expected value, choose a set of random values based on the probability and average the results N 1 E[x] ≈ x N i = Xj 1 bigger N gives better estimates throwing a die 3 rolls: 3, 1, 6 ! E[x] ≈ (3 + 1 + 6)=3 = 3.33 9 rolls: 3, 1, 6, 2, 5, 3, 4, 6, 2 ! E[x] ≈ (3 + 1 + 6 + 2 + 5 + 3 + 4 + 6 + 2)=9 = 3.51 J. Chaudhry (UNM) CS 357 19 / 41 Law of large numbers by taking N to , the error between the estimate and the expected value is statistically zero. That is, the estimate will converge to the correct value 1 N 1 P(E[x] = lim ! x ) = 1 N N i i=1 1 X This is the basic idea behind Monte Carlo estimation J. Chaudhry (UNM) CS 357 20 / 41 Continuous random variable and density function random variable: x values: x 2 [a, b] (the set of values is called sample space, which in this case is [a, b] ) b probability: density function ρ(x) > 0 with a ρ(x) dx = 1 d probability that the variable has value betweenR (c, d) ⊂ [a, b]: c ρ(x) R J. Chaudhry (UNM) CS 357 21 / 41 Uniformly distributed random variable ρ(x) is constant b a ρ(x) dx = 1 means ρ(x) = 1=(b - a) R J. Chaudhry (UNM) CS 357 22 / 41 Continuous extensions to expectation and variance Expected value b E[x] = xρ(x) dx Za b E[g(x)] = g(x)ρ(x) dx Za Variance of random variable and a function of random variable b σ2[x] = (x - E[x])2ρ(x) dx = E[x2]-(E[x])2 Za b σ2[g(x)] = (g(x)- E[g(x)])2ρ(x) dx = E[g(x)2]-(E[g(x)])2 Za Estimating the expected value of some g(x) N 1 E[g(x)] ≈ g(x ) N i Xi=1 Law of large numbers valid for continuous case too J. Chaudhry (UNM) CS 357 23 / 41 Sampling over intervals Almost all systems provide a uniform random number generator on (0, 1) Matlab: rand If we need a uniform distribution over (a, b), then we modify xk on (0, 1) by (b - a)xk + a Matlab: Generate 100 uniformly distributed values between (a, b) r = a + (b-a).*rand(100,1) Uniformly distributed integers: randi Random permutations: randperm J. Chaudhry (UNM) CS 357 24 / 41 Non-uniform distributions (Not on Exam) sampling nonuniform distributions is much more difficult if the cumulative distribution function is invertible, then we can generate the non-uniform sample from the uniform. Consider probability density function ρ(t) = λe-λt, t > 0 giving the cumulative distribution function as x x -λ -λ -λ F(x) = λe tdt = -(e t) = (1 - e x) Z0 0 thus inverting the CDF xk = - log(1 - yk)/λ where yk is uniform in (0, 1). ...not so easy in general J. Chaudhry (UNM) CS 357 25 / 41 multidementional extensions difficult domains (complex geometries) expected value E[g(x)] = g(x)ρ(x) dx Zx2Ω Ω ⊂ Rd Density function for uniform distribution: 1 ρdx = 1 =) ρ = V ZΩ where V = Ω 1 dx is the volume of Ω. J. Chaudhry (UNM)R CS 357 26 / 41 Monte Carlo Central to the PageRank (and many many other applications in fincance, science, informatics, etc) is that we randomly process something what we want to know is “on average” what is likely to happen what would happen if we have an infinite number of samples? let’s take a look at integral (a discrete limit in a sense) J.
Load more

Recommended publications
  • On Sampling from the Multivariate T Distribution by Marius Hofert
    CONTRIBUTED RESEARCH ARTICLES 129 On Sampling from the Multivariate t Distribution by Marius Hofert Abstract The multivariate normal and the multivariate t distributions belong to the most widely used multivariate distributions in statistics, quantitative risk management, and insurance. In contrast to the multivariate normal distribution, the parameterization of the multivariate t distribution does not correspond to its moments. This, paired with a non-standard implementation in the R package mvtnorm, provides traps for working with the multivariate t distribution. In this paper, common traps are clarified and corresponding recent changes to mvtnorm are presented. Introduction A supposedly simple task in statistics courses and related applications is to generate random variates from a multivariate t distribution in R. When teaching such courses, we found several fallacies one might encounter when sampling multivariate t distributions with the well-known R package mvtnorm; see Genz et al.(2013). These fallacies have recently led to improvements of the package ( ≥ 0.9-9996) which we present in this paper1. To put them in the correct context, we first address the multivariate normal distribution. The multivariate normal distribution The multivariate normal distribution can be defined in various ways, one is with its stochastic represen- tation X = m + AZ, (1) where Z = (Z1, ... , Zk) is a k-dimensional random vector with Zi, i 2 f1, ... , kg, being independent standard normal random variables, A 2 Rd×k is an (d, k)-matrix, and m 2 Rd is the mean vector. The covariance matrix of X is S = AA> and the distribution of X (that is, the d-dimensional multivariate normal distribution) is determined solely by the mean vector m and the covariance matrix S; we can thus write X ∼ Nd(m, S).
    [Show full text]
  • EFFICIENT ESTIMATION and SIMULATION of the TRUNCATED MULTIVARIATE STUDENT-T DISTRIBUTION
    Efficient estimation and simulation of the truncated multivariate student-t distribution Zdravko I. Botev, Pierre l’Ecuyer To cite this version: Zdravko I. Botev, Pierre l’Ecuyer. Efficient estimation and simulation of the truncated multivariate student-t distribution. 2015 Winter Simulation Conference, Dec 2015, Huntington Beach, United States. hal-01240154 HAL Id: hal-01240154 https://hal.inria.fr/hal-01240154 Submitted on 8 Dec 2015 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Proceedings of the 2015 Winter Simulation Conference L. Yilmaz, W. K V. Chan, I. Moon, T. M. K. Roeder, C. Macal, and M. Rosetti, eds. EFFICIENT ESTIMATION AND SIMULATION OF THE TRUNCATED MULTIVARIATE STUDENT-t DISTRIBUTION Zdravko I. Botev Pierre L’Ecuyer School of Mathematics and Statistics DIRO, Universite´ de Montreal The University of New South Wales C.P. 6128, Succ. Centre-Ville Sydney, NSW 2052, AUSTRALIA Montreal´ (Quebec),´ H3C 3J7, CANADA ABSTRACT We propose an exponential tilting method for exact simulation from the truncated multivariate student-t distribution in high dimensions as an alternative to approximate Markov Chain Monte Carlo sampling. The method also allows us to accurately estimate the probability that a random vector with multivariate student-t distribution falls in a convex polytope.
    [Show full text]
  • Stochastic Simulation APPM 7400
    Stochastic Simulation APPM 7400 Lesson 1: Random Number Generators August 27, 2018 Lesson 1: Random Number Generators Stochastic Simulation August27,2018 1/29 Random Numbers What is a random number? Lesson 1: Random Number Generators Stochastic Simulation August27,2018 2/29 Random Numbers What is a random number? “You mean like... 5?” Lesson 1: Random Number Generators Stochastic Simulation August27,2018 2/29 Random Numbers What is a random number? “You mean like... 5?” Formal definition from probability: A random number is a number uniformly distributed between 0 and 1. Lesson 1: Random Number Generators Stochastic Simulation August27,2018 2/29 Random Numbers What is a random number? “You mean like... 5?” Formal definition from probability: A random number is a number uniformly distributed between 0 and 1. (It is a realization of a uniform(0,1) random variable.) Lesson 1: Random Number Generators Stochastic Simulation August27,2018 2/29 Random Numbers Here are some realizations: 0.8763857 0.2607807 0.7060687 0.0826960 Density 0.7569021 . 0 2 4 6 8 10 0.0 0.2 0.4 0.6 0.8 1.0 sample Lesson 1: Random Number Generators Stochastic Simulation August27,2018 3/29 Random Numbers Here are some realizations: 0.8763857 0.2607807 0.7060687 0.0826960 Density 0.7569021 . 0 2 4 6 8 10 0.0 0.2 0.4 0.6 0.8 1.0 sample Lesson 1: Random Number Generators Stochastic Simulation August27,2018 3/29 Random Numbers Here are some realizations: 0.8763857 0.2607807 0.7060687 0.0826960 Density 0.7569021 .
    [Show full text]
  • Multivariate Stochastic Simulation with Subjective Multivariate Normal
    MULTIVARIATE STOCHASTIC SIMULATION WITH SUBJECTIVE MULTIVARIATE NORMAL DISTRIBUTIONS1 Peter J. Ince2 and Joseph Buongiorno3 Abstract.-In many applications of Monte Carlo simulation in forestry or forest products, it may be known that some variables are correlated. However, for simplicity, in most simulations it has been assumed that random variables are independently distributed. This report describes an alternative Monte Carlo simulation technique for subjectivelyassessed multivariate normal distributions. The method requires subjective estimates of the 99-percent confidence interval for the expected value of each random variable and of the partial correlations among the variables. The technique can be used to generate pseudorandom data corresponding to the specified distribution. If the subjective parameters do not yield a positive definite covariance matrix, the technique determines minimal adjustments in variance assumptions needed to restore positive definiteness. The method is validated and then applied to a capital investment simulation for a new papermaking technology.In that example, with ten correlated random variables, no significant difference was detected between multivariate stochastic simulation results and results that ignored the correlation. In general, however, data correlation could affect results of stochastic simulation, as shown by the validation results. INTRODUCTION MONTE CARLO TECHNIQUE Generally, a mathematical model is used in stochastic The Monte Carlo simulation technique utilizes three simulation studies. In addition to randomness, correlation may essential elements: (1) a mathematical model to calculate a exist among the variables or parameters of such models.In the discrete numerical result or outcome as a function of one or case of forest ecosystems, for example, growth can be more discrete variables, (2) a sequence of random (or influenced by correlated variables, such as temperatures and pseudorandom) numbers to represent random probabilities, and precipitations.
    [Show full text]
  • Stochastic Simulation
    Agribusiness Analysis and Forecasting Stochastic Simulation Henry Bryant Texas A&M University Henry Bryant (Texas A&M University) Agribusiness Analysis and Forecasting 1 / 19 Stochastic Simulation In economics we use simulation because we can not experiment on live subjects, a business, or the economy without injury. In other fields they can create an experiment Health sciences they feed (or treat) lots of lab rats on different chemicals to see the results. Animal science researchers feed multiple pens of steers, chickens, cows, etc. on different rations. Engineers run a motor under different controlled situations (temp, RPMs, lubricants, fuel mixes). Vets treat different pens of animals with different meds. Agronomists set up randomized block treatments for a particular seed variety with different fertilizer levels. Henry Bryant (Texas A&M University) Agribusiness Analysis and Forecasting 2 / 19 Probability Distributions Parametric and Non-Parametric Distributions Parametric Dist. have known and well defined parameters that force their shapes to known patterns. Normal Distribution - Mean and Standard Deviation. Uniform - Minimum and Maximum Bernoulli - Probability of true Beta - Alpha, Beta, Minimum, Maximum Non-Parametric Distributions do not have pre-set shapes based on known parameters. The parameters are estimated each time to make the shape of the distribution fit the data. Empirical { Actual Observations and their Probabilities. Henry Bryant (Texas A&M University) Agribusiness Analysis and Forecasting 3 / 19 Typical Problem for Risk Analysis We have a stochastic variable that needs to be included in a business model. For example: Price forecast has residuals we could not explain and they are the stochastic component we need to simulate.
    [Show full text]
  • Stochastic Vs. Deterministic Modeling of Intracellular Viral Kinetics R
    J. theor. Biol. (2002) 218, 309–321 doi:10.1006/yjtbi.3078, available online at http://www.idealibrary.com on Stochastic vs. Deterministic Modeling of Intracellular Viral Kinetics R. Srivastavawz,L.Youw,J.Summersy and J.Yinnw wDepartment of Chemical Engineering, University of Wisconsin, 3633 Engineering Hall, 1415 Engineering Drive, Madison, WI 53706, U.S.A., zMcArdle Laboratory for Cancer Research, University of Wisconsin Medical School, Madison, WI 53706, U.S.A. and yDepartment of Molecular Genetics and Microbiology, University of New Mexico School of Medicine, Albuquerque, NM 87131, U.S.A. (Received on 5 February 2002, Accepted in revised form on 3 May 2002) Within its host cell, a complex coupling of transcription, translation, genome replication, assembly, and virus release processes determines the growth rate of a virus. Mathematical models that account for these processes can provide insights into the understanding as to how the overall growth cycle depends on its constituent reactions. Deterministic models based on ordinary differential equations can capture essential relationships among virus constituents. However, an infection may be initiated by a single virus particle that delivers its genome, a single molecule of DNA or RNA, to its host cell. Under such conditions, a stochastic model that allows for inherent fluctuations in the levels of viral constituents may yield qualitatively different behavior. To compare modeling approaches, we developed a simple model of the intracellular kinetics of a generic virus, which could be implemented deterministically or stochastically. The model accounted for reactions that synthesized and depleted viral nucleic acids and structural proteins. Linear stability analysis of the deterministic model showed the existence of two nodes, one stable and one unstable.
    [Show full text]
  • Random Numbers and Stochastic Simulation
    Stochastic Simulation and Randomness Random Number Generators Quasi-Random Sequences Scientific Computing: An Introductory Survey Chapter 13 – Random Numbers and Stochastic Simulation Prof. Michael T. Heath Department of Computer Science University of Illinois at Urbana-Champaign Copyright c 2002. Reproduction permitted for noncommercial, educational use only. Michael T. Heath Scientific Computing 1 / 17 Stochastic Simulation and Randomness Random Number Generators Quasi-Random Sequences Stochastic Simulation Stochastic simulation mimics or replicates behavior of system by exploiting randomness to obtain statistical sample of possible outcomes Because of randomness involved, simulation methods are also known as Monte Carlo methods Such methods are useful for studying Nondeterministic (stochastic) processes Deterministic systems that are too complicated to model analytically Deterministic problems whose high dimensionality makes standard discretizations infeasible (e.g., Monte Carlo integration) < interactive example > < interactive example > Michael T. Heath Scientific Computing 2 / 17 Stochastic Simulation and Randomness Random Number Generators Quasi-Random Sequences Stochastic Simulation, continued Two main requirements for using stochastic simulation methods are Knowledge of relevant probability distributions Supply of random numbers for making random choices Knowledge of relevant probability distributions depends on theoretical or empirical information about physical system being simulated By simulating large number of trials, probability
    [Show full text]
  • Arxiv:2008.11456V2 [Physics.Optics]
    Efficient stochastic simulation of rate equations and photon statistics of nanolasers Emil C. Andr´e,1 Jesper Mørk,1, 2 and Martijn Wubs1,2, ∗ 1Department of Photonics Engineering, Technical University of Denmark, Ørsteds Plads 345A, DK-2800 Kgs. Lyngby, Denmark 2NanoPhoton - Center for Nanophotonics, Technical University of Denmark, Ørsteds Plads 345A, DK-2800 Kgs. Lyngby, Denmark Based on a rate equation model for single-mode two-level lasers, two algorithms for stochastically simulating the dynamics and steady-state behaviour of micro- and nanolasers are described in detail. Both methods lead to steady-state photon numbers and statistics characteristic of lasers, but one of the algorithms is shown to be significantly more efficient. This algorithm, known as Gillespie’s First Reaction Method (FRM), gives up to a thousandfold reduction in computation time compared to earlier algorithms, while also circumventing numerical issues regarding time-increment size and ordering of events. The FRM is used to examine intra-cavity photon distributions, and it is found that the numerical results follow the analytics exactly. Finally, the FRM is applied to a set of slightly altered rate equations, and it is shown that both the analytical and numerical results exhibit features that are typically associated with the presence of strong inter-emitter correlations in nanolasers. I. INTRODUCTION transitions from thermal to coherent emission, and we will investigate the prospects of introducing effectively Optical cavities on the micro- and nanometer scale can altered rates of emission in the rate equations as a way reduce the number of available modes for light emission to qualitatively describe the effects of collective inter- and increase the coupling of spontaneously emitted light emitter correlations.
    [Show full text]
  • Simulation and the Monte Carlo Method
    SIMULATION AND THE MONTE CARLO METHOD Third Edition Reuven Y. Rubinstein Technion Dirk P. Kroese University of Queensland Copyright © 2017 by John Wiley & Sons, Inc. All rights reserved. Published by John Wiley & Sons, Inc., Hoboken, New Jersey. Published simultaneously in Canada. Library of Congress Cataloging-in-Publication Data: Names: Rubinstein, Reuven Y. | Kroese, Dirk P. Title: Simulation and the Monte Carlo method. Description: Third edition / Reuven Rubinstein, Dirk P. Kroese. | Hoboken, New Jersey : John Wiley & Sons, Inc., [2017] | Series: Wiley series in probability and statistics | Includes bibliographical references and index. Identifiers: LCCN 2016020742 (print) | LCCN 2016023293 (ebook) | ISBN 9781118632161 (cloth) | ISBN 9781118632208 (pdf) | ISBN 9781118632383 (epub) Subjects: LCSH: Monte Carlo method. | Digital computer simulation. | Mathematical statistics. | Sampling (Statistics) Classification: LCC QA298 .R8 2017 (print) | LCC QA298 (ebook) | DDC 518/.282--dc23 LC record available at https://lccn.loc.gov/2016020742 Printed in the United States of America CONTENTS Preface xiii Acknowledgments xvii 1 Preliminaries 1 1.1 Introduction 1 1.2 Random Experiments 1 1.3 Conditional Probability and Independence 2 1.4 Random Variables and Probability Distributions 4 1.5 Some Important Distributions 5 1.6 Expectation 6 1.7 Joint Distributions 7 1.8 Functions of Random Variables 11 1.8.1 Linear Transformations 12 1.8.2 General Transformations 13 1.9 Transforms 14 1.10 Jointly Normal Random Variables 15 1.11 Limit Theorems 16
    [Show full text]
  • A Deterministic Monte Carlo Simulation Framework for Dam Safety Flow Control Assessment
    water Article A Deterministic Monte Carlo Simulation Framework for Dam Safety Flow Control Assessment Leanna M. King * and Slobodan P. Simonovic Department of Civil and Environmental Engineering, Western University, London, ON N6A3K7, Canada; [email protected] * Correspondence: [email protected] Received: 17 December 2019; Accepted: 5 February 2020; Published: 12 February 2020 Abstract: Simulation has become more widely applied for analysis of dam safety flow control in recent years. Stochastic simulation has proven to be a useful tool that allows for easy estimation of the overall probability of dam overtopping failure. However, it is difficult to analyze “uncommon combinations of events” with a stochastic approach given current computing abilities, because (a) the likelihood of these combinations of events is small, and (b) there may not be enough simulated instances of these rare scenarios to determine their criticality. In this research, a Deterministic Monte Carlo approach is presented, which uses an exhaustive list of possible combinations of events (scenarios) as a deterministic input. System dynamics simulation is used to model the dam system interactions so that low-level events within the system can be propagated through the model to determine high-level system outcomes. Monte Carlo iterations are performed for each input scenario. A case study is presented with results from a single example scenario to demonstrate how the simulation framework can be used to estimate the criticality parameters for each combination of events simulated. The approach can analyze these rare events in a thorough and systematic way, providing a better coverage of the possibility space as well as valuable insights into system vulnerabilities.
    [Show full text]
  • Stochastic Simulation Script for the Course in Spring 2012
    Stochastic Simulation Script for the course in spring 2012 Hansruedi K¨unsch Seminar f¨urStatistik, ETH Zurich June 2012 Contents 1 Introduction and Examples 1 1.1 What is Stochastic Simulation? . 1 1.2 Distribution of Estimators and Test Statistics . 2 1.2.1 Precision of the trimmed mean . 2 1.2.2 Bootstrap . 4 1.3 Simulation in Bayesian Statistics . 5 1.3.1 Conditional distributions of continuous random vectors . 5 1.3.2 Introduction to Bayesian statistics . 6 1.4 Simulations in Statistical Mechanics . 8 1.5 Simulations in Operations Research . 10 1.6 Simulations in Financial Mathematics . 10 1.7 The accuracy of Monte Carlo methods . 12 1.8 Other applications of stochastic simulation . 14 2 Generating Uniform Random Variables 15 2.1 Linear Congruential Generators . 16 2.2 Other Generators . 18 2.3 Combination of Generators . 19 2.4 Testing of Random Numbers . 20 3 Direct Generation of Random Variables 23 3.1 Quantile Transform . 23 3.2 Rejection Sampling . 24 3.3 Ratio of Uniform Random Variables . 26 3.4 Relations between Distributions . 27 3.4.1 The Normal Distribution . 27 3.4.2 The Poisson Distribution . 29 3.5 Summary: Simulation of Selected Distributions . 29 3.6 Random Sampling and Random Permutations . 31 3.7 Importance Sampling . 32 3.8 Markov Chains and Markov Processes . 32 3.8.1 Simulation of Stochastic Differential Equations . 33 3.9 Variance Reduction . 35 3.9.1 Antithetic Variates . 35 3.9.2 Control Variates . 36 3.9.3 Importance Sampling and Variance Reduction . 37 3.9.4 Quasi-Monte Carlo .
    [Show full text]
  • Statistics and Simulation
    Statistics and Simulation By Smita Skrivanek, Principal Statistician, MoreSteam.com LLC Stochastic Simulation Stochastic simulation has played an increasingly large role in statistics with rapid increases in computing power. The exercises in the MoreSteam courses are examples of simulation. Simulations are used to model complicated processes, estimate distributions of estimators (using methods such as bootstrap), and have dramatically increased the use of an entire field of statistics, Bayesian statistics, since simulation can be used to estimate the distribution of its statistics, which are usually analytically intractable except in the simplest situations. So what is ‘stochastic simulation’? Stochastic simulation, a.k.a Monte Carlo simulation, is the use of random number generators to produce data from a mathematical model. A random number generator is any process that produces data whose observations are independent and identically distributed (i.i.d) from some distribution. It is a method used to evaluate processes or systems using probabilities based on sampling repeatedly from randomly generated numbers. Simulation by random sampling was in use during the genesis of probability theory, long before computers. Toss a coin a hundred times and count the number of times it falls Heads. Dividing the total number of times the coin falls Heads by the total number of tosses gives you an estimate of the probability of Heads for that coin, using simulation. The error associated with that estimate is called “Monte Carlo error”. Stochastic simulation was used as far back as in 1777 by Buffon to estimate the probability of a needle falling across a line in a uniform grid of parallel lines, known as Buffon’s needle problem.
    [Show full text]