random-variable-stream.cc

Go to the documentation of this file.

/*
 * Copyright (c) 2006 Georgia Tech Research Corporation
 * Copyright (c) 2011 Mathieu Lacage
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation;
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 * Authors: Rajib Bhattacharjea<raj.b@gatech.edu>
 *          Hadi Arbabi<marbabi@cs.odu.edu>
 *          Mathieu Lacage <mathieu.lacage@gmail.com>
 *
 * Modified by Mitch Watrous <watrous@u.washington.edu>
 *
 */
#include "random-variable-stream.h"
 
#include "assert.h"
#include "boolean.h"
#include "double.h"
#include "integer.h"
#include "log.h"
#include "pointer.h"
#include "rng-seed-manager.h"
#include "rng-stream.h"
#include "string.h"
 
#include <algorithm> // upper_bound
#include <cmath>
#include <iostream>
 
namespace ns3
{
 
NS_LOG_COMPONENT_DEFINE("RandomVariableStream");
 
NS_OBJECT_ENSURE_REGISTERED(RandomVariableStream);
 
TypeId
RandomVariableStream::GetTypeId()
{
    static TypeId tid = TypeId("ns3::RandomVariableStream")
                            .SetParent<Object>()
                            .SetGroupName("Core")
                            .AddAttribute("Stream",
                                          "The stream number for this RNG stream. -1 means "
                                          "\"allocate a stream automatically\". "
                                          "Note that if -1 is set, Get will return -1 so that it "
                                          "is not possible to know which "
                                          "value was automatically allocated.",
                                          IntegerValue(-1),
                                          MakeIntegerAccessor(&RandomVariableStream::SetStream,
                                                              &RandomVariableStream::GetStream),
                                          MakeIntegerChecker<int64_t>())
                            .AddAttribute("Antithetic",
                                          "Set this RNG stream to generate antithetic values",
                                          BooleanValue(false),
                                          MakeBooleanAccessor(&RandomVariableStream::SetAntithetic,
                                                              &RandomVariableStream::IsAntithetic),
                                          MakeBooleanChecker());
    return tid;
}
 
RandomVariableStream::RandomVariableStream()
    : m_rng(nullptr)
{
    NS_LOG_FUNCTION(this);
}
 
RandomVariableStream::~RandomVariableStream()
{
    NS_LOG_FUNCTION(this);
    delete m_rng;
}
 
void
RandomVariableStream::SetAntithetic(bool isAntithetic)
{
    NS_LOG_FUNCTION(this << isAntithetic);
    m_isAntithetic = isAntithetic;
}
 
bool
RandomVariableStream::IsAntithetic() const
{
    NS_LOG_FUNCTION(this);
    return m_isAntithetic;
}
 
uint32_t
RandomVariableStream::GetInteger()
{
    NS_LOG_FUNCTION(this);
    return static_cast<uint32_t>(GetValue());
}
 
void
RandomVariableStream::SetStream(int64_t stream)
{
    NS_LOG_FUNCTION(this << stream);
    // negative values are not legal.
    NS_ASSERT(stream >= -1);
    delete m_rng;
    if (stream == -1)
    {
        // The first 2^63 streams are reserved for automatic stream
        // number assignment.
        uint64_t nextStream = RngSeedManager::GetNextStreamIndex();
        NS_ASSERT(nextStream <= ((1ULL) << 63));
        m_rng = new RngStream(RngSeedManager::GetSeed(), nextStream, RngSeedManager::GetRun());
    }
    else
    {
        // The last 2^63 streams are reserved for deterministic stream
        // number assignment.
        uint64_t base = ((1ULL) << 63);
        uint64_t target = base + stream;
        m_rng = new RngStream(RngSeedManager::GetSeed(), target, RngSeedManager::GetRun());
    }
    m_stream = stream;
}
 
int64_t
RandomVariableStream::GetStream() const
{
    NS_LOG_FUNCTION(this);
    return m_stream;
}
 
RngStream*
RandomVariableStream::Peek() const
{
    NS_LOG_FUNCTION(this);
    return m_rng;
}
 
NS_OBJECT_ENSURE_REGISTERED(UniformRandomVariable);
 
TypeId
UniformRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::UniformRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<UniformRandomVariable>()
            .AddAttribute("Min",
                          "The lower bound on the values returned by this RNG stream.",
                          DoubleValue(0),
                          MakeDoubleAccessor(&UniformRandomVariable::m_min),
                          MakeDoubleChecker<double>())
            .AddAttribute("Max",
                          "The upper bound on the values returned by this RNG stream.",
                          DoubleValue(1.0),
                          MakeDoubleAccessor(&UniformRandomVariable::m_max),
                          MakeDoubleChecker<double>());
    return tid;
}
 
UniformRandomVariable::UniformRandomVariable()
{
    // m_min and m_max are initialized after constructor by attributes
    NS_LOG_FUNCTION(this);
}
 
double
UniformRandomVariable::GetMin() const
{
    NS_LOG_FUNCTION(this);
    return m_min;
}
 
double
UniformRandomVariable::GetMax() const
{
    NS_LOG_FUNCTION(this);
    return m_max;
}
 
double
UniformRandomVariable::GetValue(double min, double max)
{
    NS_LOG_FUNCTION(this << min << max);
    double v = min + Peek()->RandU01() * (max - min);
    if (IsAntithetic())
    {
        v = min + (max - v);
    }
    return v;
}
 
uint32_t
UniformRandomVariable::GetInteger(uint32_t min, uint32_t max)
{
    NS_LOG_FUNCTION(this << min << max);
    NS_ASSERT(min <= max);
    return static_cast<uint32_t>(GetValue((double)(min), (double)(max) + 1.0));
}
 
double
UniformRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_min, m_max);
}
 
uint32_t
UniformRandomVariable::GetInteger()
{
    NS_LOG_FUNCTION(this);
    return static_cast<uint32_t>(GetValue(m_min, m_max + 1));
}
 
NS_OBJECT_ENSURE_REGISTERED(ConstantRandomVariable);
 
TypeId
ConstantRandomVariable::GetTypeId()
{
    static TypeId tid = TypeId("ns3::ConstantRandomVariable")
                            .SetParent<RandomVariableStream>()
                            .SetGroupName("Core")
                            .AddConstructor<ConstantRandomVariable>()
                            .AddAttribute("Constant",
                                          "The constant value returned by this RNG stream.",
                                          DoubleValue(0),
                                          MakeDoubleAccessor(&ConstantRandomVariable::m_constant),
                                          MakeDoubleChecker<double>());
    return tid;
}
 
ConstantRandomVariable::ConstantRandomVariable()
{
    // m_constant is initialized after constructor by attributes
    NS_LOG_FUNCTION(this);
}
 
double
ConstantRandomVariable::GetConstant() const
{
    NS_LOG_FUNCTION(this);
    return m_constant;
}
 
double
ConstantRandomVariable::GetValue(double constant)
{
    NS_LOG_FUNCTION(this << constant);
    return constant;
}
 
uint32_t
ConstantRandomVariable::GetInteger(uint32_t constant)
{
    NS_LOG_FUNCTION(this << constant);
    return constant;
}
 
double
ConstantRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_constant);
}
 
NS_OBJECT_ENSURE_REGISTERED(SequentialRandomVariable);
 
TypeId
SequentialRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::SequentialRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<SequentialRandomVariable>()
            .AddAttribute("Min",
                          "The first value of the sequence.",
                          DoubleValue(0),
                          MakeDoubleAccessor(&SequentialRandomVariable::m_min),
                          MakeDoubleChecker<double>())
            .AddAttribute("Max",
                          "One more than the last value of the sequence.",
                          DoubleValue(0),
                          MakeDoubleAccessor(&SequentialRandomVariable::m_max),
                          MakeDoubleChecker<double>())
            .AddAttribute("Increment",
                          "The sequence random variable increment.",
                          StringValue("ns3::ConstantRandomVariable[Constant=1]"),
                          MakePointerAccessor(&SequentialRandomVariable::m_increment),
                          MakePointerChecker<RandomVariableStream>())
            .AddAttribute("Consecutive",
                          "The number of times each member of the sequence is repeated.",
                          IntegerValue(1),
                          MakeIntegerAccessor(&SequentialRandomVariable::m_consecutive),
                          MakeIntegerChecker<uint32_t>());
    return tid;
}
 
SequentialRandomVariable::SequentialRandomVariable()
    : m_current(0),
      m_currentConsecutive(0),
      m_isCurrentSet(false)
{
    // m_min, m_max, m_increment, and m_consecutive are initialized
    // after constructor by attributes.
    NS_LOG_FUNCTION(this);
}
 
double
SequentialRandomVariable::GetMin() const
{
    NS_LOG_FUNCTION(this);
    return m_min;
}
 
double
SequentialRandomVariable::GetMax() const
{
    NS_LOG_FUNCTION(this);
    return m_max;
}
 
Ptr<RandomVariableStream>
SequentialRandomVariable::GetIncrement() const
{
    NS_LOG_FUNCTION(this);
    return m_increment;
}
 
uint32_t
SequentialRandomVariable::GetConsecutive() const
{
    NS_LOG_FUNCTION(this);
    return m_consecutive;
}
 
double
SequentialRandomVariable::GetValue()
{
    // Set the current sequence value if it hasn't been set.
    NS_LOG_FUNCTION(this);
    if (!m_isCurrentSet)
    {
        // Start the sequence at its minimum value.
        m_current = m_min;
        m_isCurrentSet = true;
    }
 
    // Return a sequential series of values
    double r = m_current;
    if (++m_currentConsecutive == m_consecutive)
    { // Time to advance to next
        m_currentConsecutive = 0;
        m_current += m_increment->GetValue();
        if (m_current >= m_max)
        {
            m_current = m_min + (m_current - m_max);
        }
    }
    return r;
}
 
NS_OBJECT_ENSURE_REGISTERED(ExponentialRandomVariable);
 
TypeId
ExponentialRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::ExponentialRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<ExponentialRandomVariable>()
            .AddAttribute("Mean",
                          "The mean of the values returned by this RNG stream.",
                          DoubleValue(1.0),
                          MakeDoubleAccessor(&ExponentialRandomVariable::m_mean),
                          MakeDoubleChecker<double>())
            .AddAttribute("Bound",
                          "The upper bound on the values returned by this RNG stream.",
                          DoubleValue(0.0),
                          MakeDoubleAccessor(&ExponentialRandomVariable::m_bound),
                          MakeDoubleChecker<double>());
    return tid;
}
 
ExponentialRandomVariable::ExponentialRandomVariable()
{
    // m_mean and m_bound are initialized after constructor by attributes
    NS_LOG_FUNCTION(this);
}
 
double
ExponentialRandomVariable::GetMean() const
{
    NS_LOG_FUNCTION(this);
    return m_mean;
}
 
double
ExponentialRandomVariable::GetBound() const
{
    NS_LOG_FUNCTION(this);
    return m_bound;
}
 
double
ExponentialRandomVariable::GetValue(double mean, double bound)
{
    NS_LOG_FUNCTION(this << mean << bound);
    while (1)
    {
        // Get a uniform random variable in [0,1].
        double v = Peek()->RandU01();
        if (IsAntithetic())
        {
            v = (1 - v);
        }
 
        // Calculate the exponential random variable.
        double r = -mean * std::log(v);
 
        // Use this value if it's acceptable.
        if (bound == 0 || r <= bound)
        {
            return r;
        }
    }
}
 
uint32_t
ExponentialRandomVariable::GetInteger(uint32_t mean, uint32_t bound)
{
    NS_LOG_FUNCTION(this << mean << bound);
    return static_cast<uint32_t>(GetValue(mean, bound));
}
 
double
ExponentialRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_mean, m_bound);
}
 
NS_OBJECT_ENSURE_REGISTERED(ParetoRandomVariable);
 
TypeId
ParetoRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::ParetoRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<ParetoRandomVariable>()
            .AddAttribute(
                "Scale",
                "The scale parameter for the Pareto distribution returned by this RNG stream.",
                DoubleValue(1.0),
                MakeDoubleAccessor(&ParetoRandomVariable::m_scale),
                MakeDoubleChecker<double>())
            .AddAttribute(
                "Shape",
                "The shape parameter for the Pareto distribution returned by this RNG stream.",
                DoubleValue(2.0),
                MakeDoubleAccessor(&ParetoRandomVariable::m_shape),
                MakeDoubleChecker<double>())
            .AddAttribute(
                "Bound",
                "The upper bound on the values returned by this RNG stream (if non-zero).",
                DoubleValue(0.0),
                MakeDoubleAccessor(&ParetoRandomVariable::m_bound),
                MakeDoubleChecker<double>());
    return tid;
}
 
ParetoRandomVariable::ParetoRandomVariable()
{
    // m_shape, m_shape, and m_bound are initialized after constructor
    // by attributes
    NS_LOG_FUNCTION(this);
}
 
double
ParetoRandomVariable::GetScale() const
{
    NS_LOG_FUNCTION(this);
    return m_scale;
}
 
double
ParetoRandomVariable::GetShape() const
{
    NS_LOG_FUNCTION(this);
    return m_shape;
}
 
double
ParetoRandomVariable::GetBound() const
{
    NS_LOG_FUNCTION(this);
    return m_bound;
}
 
double
ParetoRandomVariable::GetValue(double scale, double shape, double bound)
{
    // Calculate the scale parameter.
    NS_LOG_FUNCTION(this << scale << shape << bound);
 
    while (1)
    {
        // Get a uniform random variable in [0,1].
        double v = Peek()->RandU01();
        if (IsAntithetic())
        {
            v = (1 - v);
        }
 
        // Calculate the Pareto random variable.
        double r = (scale * (1.0 / std::pow(v, 1.0 / shape)));
 
        // Use this value if it's acceptable.
        if (bound == 0 || r <= bound)
        {
            return r;
        }
    }
}
 
uint32_t
ParetoRandomVariable::GetInteger(uint32_t scale, uint32_t shape, uint32_t bound)
{
    NS_LOG_FUNCTION(this << scale << shape << bound);
    return static_cast<uint32_t>(GetValue(scale, shape, bound));
}
 
double
ParetoRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_scale, m_shape, m_bound);
}
 
NS_OBJECT_ENSURE_REGISTERED(WeibullRandomVariable);
 
TypeId
WeibullRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::WeibullRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<WeibullRandomVariable>()
            .AddAttribute(
                "Scale",
                "The scale parameter for the Weibull distribution returned by this RNG stream.",
                DoubleValue(1.0),
                MakeDoubleAccessor(&WeibullRandomVariable::m_scale),
                MakeDoubleChecker<double>())
            .AddAttribute(
                "Shape",
                "The shape parameter for the Weibull distribution returned by this RNG stream.",
                DoubleValue(1),
                MakeDoubleAccessor(&WeibullRandomVariable::m_shape),
                MakeDoubleChecker<double>())
            .AddAttribute("Bound",
                          "The upper bound on the values returned by this RNG stream.",
                          DoubleValue(0.0),
                          MakeDoubleAccessor(&WeibullRandomVariable::m_bound),
                          MakeDoubleChecker<double>());
    return tid;
}
 
WeibullRandomVariable::WeibullRandomVariable()
{
    // m_scale, m_shape, and m_bound are initialized after constructor
    // by attributes
    NS_LOG_FUNCTION(this);
}
 
double
WeibullRandomVariable::GetScale() const
{
    NS_LOG_FUNCTION(this);
    return m_scale;
}
 
double
WeibullRandomVariable::GetShape() const
{
    NS_LOG_FUNCTION(this);
    return m_shape;
}
 
double
WeibullRandomVariable::GetBound() const
{
    NS_LOG_FUNCTION(this);
    return m_bound;
}
 
double
WeibullRandomVariable::GetValue(double scale, double shape, double bound)
{
    NS_LOG_FUNCTION(this << scale << shape << bound);
    double exponent = 1.0 / shape;
    while (1)
    {
        // Get a uniform random variable in [0,1].
        double v = Peek()->RandU01();
        if (IsAntithetic())
        {
            v = (1 - v);
        }
 
        // Calculate the Weibull random variable.
        double r = scale * std::pow(-std::log(v), exponent);
 
        // Use this value if it's acceptable.
        if (bound == 0 || r <= bound)
        {
            return r;
        }
    }
}
 
uint32_t
WeibullRandomVariable::GetInteger(uint32_t scale, uint32_t shape, uint32_t bound)
{
    NS_LOG_FUNCTION(this << scale << shape << bound);
    return static_cast<uint32_t>(GetValue(scale, shape, bound));
}
 
double
WeibullRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_scale, m_shape, m_bound);
}
 
NS_OBJECT_ENSURE_REGISTERED(NormalRandomVariable);
 
const double NormalRandomVariable::INFINITE_VALUE = 1e307;
 
TypeId
NormalRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::NormalRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<NormalRandomVariable>()
            .AddAttribute("Mean",
                          "The mean value for the normal distribution returned by this RNG stream.",
                          DoubleValue(0.0),
                          MakeDoubleAccessor(&NormalRandomVariable::m_mean),
                          MakeDoubleChecker<double>())
            .AddAttribute(
                "Variance",
                "The variance value for the normal distribution returned by this RNG stream.",
                DoubleValue(1.0),
                MakeDoubleAccessor(&NormalRandomVariable::m_variance),
                MakeDoubleChecker<double>())
            .AddAttribute("Bound",
                          "The bound on the values returned by this RNG stream.",
                          DoubleValue(INFINITE_VALUE),
                          MakeDoubleAccessor(&NormalRandomVariable::m_bound),
                          MakeDoubleChecker<double>());
    return tid;
}
 
NormalRandomVariable::NormalRandomVariable()
    : m_nextValid(false)
{
    // m_mean, m_variance, and m_bound are initialized after constructor
    // by attributes
    NS_LOG_FUNCTION(this);
}
 
double
NormalRandomVariable::GetMean() const
{
    NS_LOG_FUNCTION(this);
    return m_mean;
}
 
double
NormalRandomVariable::GetVariance() const
{
    NS_LOG_FUNCTION(this);
    return m_variance;
}
 
double
NormalRandomVariable::GetBound() const
{
    NS_LOG_FUNCTION(this);
    return m_bound;
}
 
double
NormalRandomVariable::GetValue(double mean, double variance, double bound)
{
    NS_LOG_FUNCTION(this << mean << variance << bound);
    if (m_nextValid)
    { // use previously generated
        m_nextValid = false;
        double x2 = mean + m_v2 * m_y * std::sqrt(variance);
        if (std::fabs(x2 - mean) <= bound)
        {
            return x2;
        }
    }
    while (1)
    { // See Simulation Modeling and Analysis p. 466 (Averill Law)
        // for algorithm; basically a Box-Muller transform:
        // http://en.wikipedia.org/wiki/Box-Muller_transform
        double u1 = Peek()->RandU01();
        double u2 = Peek()->RandU01();
        if (IsAntithetic())
        {
            u1 = (1 - u1);
            u2 = (1 - u2);
        }
        double v1 = 2 * u1 - 1;
        double v2 = 2 * u2 - 1;
        double w = v1 * v1 + v2 * v2;
        if (w <= 1.0)
        { // Got good pair
            double y = std::sqrt((-2 * std::log(w)) / w);
            double x1 = mean + v1 * y * std::sqrt(variance);
            // if x1 is in bounds, return it, cache v2 and y
            if (std::fabs(x1 - mean) <= bound)
            {
                m_nextValid = true;
                m_y = y;
                m_v2 = v2;
                return x1;
            }
            // otherwise try and return the other if it is valid
            double x2 = mean + v2 * y * std::sqrt(variance);
            if (std::fabs(x2 - mean) <= bound)
            {
                m_nextValid = false;
                return x2;
            }
            // otherwise, just run this loop again
        }
    }
}
 
uint32_t
NormalRandomVariable::GetInteger(uint32_t mean, uint32_t variance, uint32_t bound)
{
    NS_LOG_FUNCTION(this << mean << variance << bound);
    return static_cast<uint32_t>(GetValue(mean, variance, bound));
}
 
double
NormalRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_mean, m_variance, m_bound);
}
 
NS_OBJECT_ENSURE_REGISTERED(LogNormalRandomVariable);
 
TypeId
LogNormalRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::LogNormalRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<LogNormalRandomVariable>()
            .AddAttribute(
                "Mu",
                "The mu value for the log-normal distribution returned by this RNG stream.",
                DoubleValue(0.0),
                MakeDoubleAccessor(&LogNormalRandomVariable::m_mu),
                MakeDoubleChecker<double>())
            .AddAttribute(
                "Sigma",
                "The sigma value for the log-normal distribution returned by this RNG stream.",
                DoubleValue(1.0),
                MakeDoubleAccessor(&LogNormalRandomVariable::m_sigma),
                MakeDoubleChecker<double>());
    return tid;
}
 
LogNormalRandomVariable::LogNormalRandomVariable()
    : m_nextValid(false)
{
    // m_mu and m_sigma are initialized after constructor by
    // attributes
    NS_LOG_FUNCTION(this);
}
 
double
LogNormalRandomVariable::GetMu() const
{
    NS_LOG_FUNCTION(this);
    return m_mu;
}
 
double
LogNormalRandomVariable::GetSigma() const
{
    NS_LOG_FUNCTION(this);
    return m_sigma;
}
 
// The code from this function was adapted from the GNU Scientific
// Library 1.8:
/* randist/lognormal.c
 *
 * Copyright (C) 1996, 1997, 1998, 1999, 2000 James Theiler, Brian Gough
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or (at
 * your option) any later version.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
 */
/* The lognormal distribution has the form
 
   p(x) dx = 1/(x * sqrt(2 pi sigma^2)) exp(-(ln(x) - zeta)^2/2 sigma^2) dx
 
   for x > 0. Lognormal random numbers are the exponentials of
   gaussian random numbers */
double
LogNormalRandomVariable::GetValue(double mu, double sigma)
{
    if (m_nextValid)
    { // use previously generated
        m_nextValid = false;
        double normal = m_v2 * m_normal;
 
        return std::exp(sigma * normal + mu);
    }
 
    double v1;
    double v2;
    double r2;
    double normal;
    double x;
 
    NS_LOG_FUNCTION(this << mu << sigma);
 
    do
    {
        /* choose x,y in uniform square (-1,-1) to (+1,+1) */
 
        double u1 = Peek()->RandU01();
        double u2 = Peek()->RandU01();
        if (IsAntithetic())
        {
            u1 = (1 - u1);
            u2 = (1 - u2);
        }
 
        v1 = -1 + 2 * u1;
        v2 = -1 + 2 * u2;
 
        /* see if it is in the unit circle */
        r2 = v1 * v1 + v2 * v2;
    } while (r2 > 1.0 || r2 == 0);
 
    m_normal = std::sqrt(-2.0 * std::log(r2) / r2);
    normal = v1 * m_normal;
    m_nextValid = true;
    m_v2 = v2;
 
    x = std::exp(sigma * normal + mu);
 
    return x;
}
 
uint32_t
LogNormalRandomVariable::GetInteger(uint32_t mu, uint32_t sigma)
{
    NS_LOG_FUNCTION(this << mu << sigma);
    return static_cast<uint32_t>(GetValue(mu, sigma));
}
 
double
LogNormalRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_mu, m_sigma);
}
 
NS_OBJECT_ENSURE_REGISTERED(GammaRandomVariable);
 
TypeId
GammaRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::GammaRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<GammaRandomVariable>()
            .AddAttribute("Alpha",
                          "The alpha value for the gamma distribution returned by this RNG stream.",
                          DoubleValue(1.0),
                          MakeDoubleAccessor(&GammaRandomVariable::m_alpha),
                          MakeDoubleChecker<double>())
            .AddAttribute("Beta",
                          "The beta value for the gamma distribution returned by this RNG stream.",
                          DoubleValue(1.0),
                          MakeDoubleAccessor(&GammaRandomVariable::m_beta),
                          MakeDoubleChecker<double>());
    return tid;
}
 
GammaRandomVariable::GammaRandomVariable()
    : m_nextValid(false)
{
    // m_alpha and m_beta are initialized after constructor by
    // attributes
    NS_LOG_FUNCTION(this);
}
 
double
GammaRandomVariable::GetAlpha() const
{
    NS_LOG_FUNCTION(this);
    return m_alpha;
}
 
double
GammaRandomVariable::GetBeta() const
{
    NS_LOG_FUNCTION(this);
    return m_beta;
}
 
/*
  The code for the following generator functions was adapted from ns-2
  tools/ranvar.cc
 
  Originally the algorithm was devised by Marsaglia in 2000:
  G. Marsaglia, W. W. Tsang: A simple method for generating Gamma variables
  ACM Transactions on mathematical software, Vol. 26, No. 3, Sept. 2000
 
  The Gamma distribution density function has the form
 
                             x^(alpha-1) * exp(-x/beta)
        p(x; alpha, beta) = ----------------------------
                             beta^alpha * Gamma(alpha)
 
  for x > 0.
*/
double
GammaRandomVariable::GetValue(double alpha, double beta)
{
    NS_LOG_FUNCTION(this << alpha << beta);
    if (alpha < 1)
    {
        double u = Peek()->RandU01();
        if (IsAntithetic())
        {
            u = (1 - u);
        }
        return GetValue(1.0 + alpha, beta) * std::pow(u, 1.0 / alpha);
    }
 
    double x;
    double v;
    double u;
    double d = alpha - 1.0 / 3.0;
    double c = (1.0 / 3.0) / std::sqrt(d);
 
    while (1)
    {
        do
        {
            // Get a value from a normal distribution that has mean
            // zero, variance 1, and no bound.
            double mean = 0.0;
            double variance = 1.0;
            double bound = NormalRandomVariable::INFINITE_VALUE;
            x = GetNormalValue(mean, variance, bound);
 
            v = 1.0 + c * x;
        } while (v <= 0);
 
        v = v * v * v;
        u = Peek()->RandU01();
        if (IsAntithetic())
        {
            u = (1 - u);
        }
        if (u < 1 - 0.0331 * x * x * x * x)
        {
            break;
        }
        if (std::log(u) < 0.5 * x * x + d * (1 - v + std::log(v)))
        {
            break;
        }
    }
 
    return beta * d * v;
}
 
double
GammaRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_alpha, m_beta);
}
 
double
GammaRandomVariable::GetNormalValue(double mean, double variance, double bound)
{
    NS_LOG_FUNCTION(this << mean << variance << bound);
    if (m_nextValid)
    { // use previously generated
        m_nextValid = false;
        double x2 = mean + m_v2 * m_y * std::sqrt(variance);
        if (std::fabs(x2 - mean) <= bound)
        {
            return x2;
        }
    }
    while (1)
    { // See Simulation Modeling and Analysis p. 466 (Averill Law)
        // for algorithm; basically a Box-Muller transform:
        // http://en.wikipedia.org/wiki/Box-Muller_transform
        double u1 = Peek()->RandU01();
        double u2 = Peek()->RandU01();
        if (IsAntithetic())
        {
            u1 = (1 - u1);
            u2 = (1 - u2);
        }
        double v1 = 2 * u1 - 1;
        double v2 = 2 * u2 - 1;
        double w = v1 * v1 + v2 * v2;
        if (w <= 1.0)
        { // Got good pair
            double y = std::sqrt((-2 * std::log(w)) / w);
            double x1 = mean + v1 * y * std::sqrt(variance);
            // if x1 is in bounds, return it, cache v2 an y
            if (std::fabs(x1 - mean) <= bound)
            {
                m_nextValid = true;
                m_y = y;
                m_v2 = v2;
                return x1;
            }
            // otherwise try and return the other if it is valid
            double x2 = mean + v2 * y * std::sqrt(variance);
            if (std::fabs(x2 - mean) <= bound)
            {
                m_nextValid = false;
                return x2;
            }
            // otherwise, just run this loop again
        }
    }
}
 
NS_OBJECT_ENSURE_REGISTERED(ErlangRandomVariable);
 
TypeId
ErlangRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::ErlangRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<ErlangRandomVariable>()
            .AddAttribute("K",
                          "The k value for the Erlang distribution returned by this RNG stream.",
                          IntegerValue(1),
                          MakeIntegerAccessor(&ErlangRandomVariable::m_k),
                          MakeIntegerChecker<uint32_t>())
            .AddAttribute(
                "Lambda",
                "The lambda value for the Erlang distribution returned by this RNG stream.",
                DoubleValue(1.0),
                MakeDoubleAccessor(&ErlangRandomVariable::m_lambda),
                MakeDoubleChecker<double>());
    return tid;
}
 
ErlangRandomVariable::ErlangRandomVariable()
{
    // m_k and m_lambda are initialized after constructor by attributes
    NS_LOG_FUNCTION(this);
}
 
uint32_t
ErlangRandomVariable::GetK() const
{
    NS_LOG_FUNCTION(this);
    return m_k;
}
 
double
ErlangRandomVariable::GetLambda() const
{
    NS_LOG_FUNCTION(this);
    return m_lambda;
}
 
/*
  The code for the following generator functions was adapted from ns-2
  tools/ranvar.cc
 
  The Erlang distribution density function has the form
 
                           x^(k-1) * exp(-x/lambda)
        p(x; k, lambda) = ---------------------------
                             lambda^k * (k-1)!
 
  for x > 0.
*/
double
ErlangRandomVariable::GetValue(uint32_t k, double lambda)
{
    NS_LOG_FUNCTION(this << k << lambda);
    double mean = lambda;
    double bound = 0.0;
 
    double result = 0;
    for (unsigned int i = 0; i < k; ++i)
    {
        result += GetExponentialValue(mean, bound);
    }
 
    return result;
}
 
uint32_t
ErlangRandomVariable::GetInteger(uint32_t k, uint32_t lambda)
{
    NS_LOG_FUNCTION(this << k << lambda);
    return static_cast<uint32_t>(GetValue(k, lambda));
}
 
double
ErlangRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_k, m_lambda);
}
 
double
ErlangRandomVariable::GetExponentialValue(double mean, double bound)
{
    NS_LOG_FUNCTION(this << mean << bound);
    while (1)
    {
        // Get a uniform random variable in [0,1].
        double v = Peek()->RandU01();
        if (IsAntithetic())
        {
            v = (1 - v);
        }
 
        // Calculate the exponential random variable.
        double r = -mean * std::log(v);
 
        // Use this value if it's acceptable.
        if (bound == 0 || r <= bound)
        {
            return r;
        }
    }
}
 
NS_OBJECT_ENSURE_REGISTERED(TriangularRandomVariable);
 
TypeId
TriangularRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::TriangularRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<TriangularRandomVariable>()
            .AddAttribute(
                "Mean",
                "The mean value for the triangular distribution returned by this RNG stream.",
                DoubleValue(0.5),
                MakeDoubleAccessor(&TriangularRandomVariable::m_mean),
                MakeDoubleChecker<double>())
            .AddAttribute("Min",
                          "The lower bound on the values returned by this RNG stream.",
                          DoubleValue(0.0),
                          MakeDoubleAccessor(&TriangularRandomVariable::m_min),
                          MakeDoubleChecker<double>())
            .AddAttribute("Max",
                          "The upper bound on the values returned by this RNG stream.",
                          DoubleValue(1.0),
                          MakeDoubleAccessor(&TriangularRandomVariable::m_max),
                          MakeDoubleChecker<double>());
    return tid;
}
 
TriangularRandomVariable::TriangularRandomVariable()
{
    // m_mean, m_min, and m_max are initialized after constructor by
    // attributes
    NS_LOG_FUNCTION(this);
}
 
double
TriangularRandomVariable::GetMean() const
{
    NS_LOG_FUNCTION(this);
    return m_mean;
}
 
double
TriangularRandomVariable::GetMin() const
{
    NS_LOG_FUNCTION(this);
    return m_min;
}
 
double
TriangularRandomVariable::GetMax() const
{
    NS_LOG_FUNCTION(this);
    return m_max;
}
 
double
TriangularRandomVariable::GetValue(double mean, double min, double max)
{
    // Calculate the mode.
    NS_LOG_FUNCTION(this << mean << min << max);
    double mode = 3.0 * mean - min - max;
 
    // Get a uniform random variable in [0,1].
    double u = Peek()->RandU01();
    if (IsAntithetic())
    {
        u = (1 - u);
    }
 
    // Calculate the triangular random variable.
    if (u <= (mode - min) / (max - min))
    {
        return min + std::sqrt(u * (max - min) * (mode - min));
    }
    else
    {
        return max - std::sqrt((1 - u) * (max - min) * (max - mode));
    }
}
 
uint32_t
TriangularRandomVariable::GetInteger(uint32_t mean, uint32_t min, uint32_t max)
{
    NS_LOG_FUNCTION(this << mean << min << max);
    return static_cast<uint32_t>(GetValue(mean, min, max));
}
 
double
TriangularRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_mean, m_min, m_max);
}
 
NS_OBJECT_ENSURE_REGISTERED(ZipfRandomVariable);
 
TypeId
ZipfRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::ZipfRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<ZipfRandomVariable>()
            .AddAttribute("N",
                          "The n value for the Zipf distribution returned by this RNG stream.",
                          IntegerValue(1),
                          MakeIntegerAccessor(&ZipfRandomVariable::m_n),
                          MakeIntegerChecker<uint32_t>())
            .AddAttribute("Alpha",
                          "The alpha value for the Zipf distribution returned by this RNG stream.",
                          DoubleValue(0.0),
                          MakeDoubleAccessor(&ZipfRandomVariable::m_alpha),
                          MakeDoubleChecker<double>());
    return tid;
}
 
ZipfRandomVariable::ZipfRandomVariable()
{
    // m_n and m_alpha are initialized after constructor by attributes
    NS_LOG_FUNCTION(this);
}
 
uint32_t
ZipfRandomVariable::GetN() const
{
    NS_LOG_FUNCTION(this);
    return m_n;
}
 
double
ZipfRandomVariable::GetAlpha() const
{
    NS_LOG_FUNCTION(this);
    return m_alpha;
}
 
double
ZipfRandomVariable::GetValue(uint32_t n, double alpha)
{
    NS_LOG_FUNCTION(this << n << alpha);
    // Calculate the normalization constant c.
    m_c = 0.0;
    for (uint32_t i = 1; i <= n; i++)
    {
        m_c += (1.0 / std::pow((double)i, alpha));
    }
    m_c = 1.0 / m_c;
 
    // Get a uniform random variable in [0,1].
    double u = Peek()->RandU01();
    if (IsAntithetic())
    {
        u = (1 - u);
    }
 
    double sum_prob = 0;
    double zipf_value = 0;
    for (uint32_t i = 1; i <= m_n; i++)
    {
        sum_prob += m_c / std::pow((double)i, m_alpha);
        if (sum_prob > u)
        {
            zipf_value = i;
            break;
        }
    }
    return zipf_value;
}
 
uint32_t
ZipfRandomVariable::GetInteger(uint32_t n, uint32_t alpha)
{
    NS_LOG_FUNCTION(this << n << alpha);
    return static_cast<uint32_t>(GetValue(n, alpha));
}
 
double
ZipfRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_n, m_alpha);
}
 
NS_OBJECT_ENSURE_REGISTERED(ZetaRandomVariable);
 
TypeId
ZetaRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::ZetaRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<ZetaRandomVariable>()
            .AddAttribute("Alpha",
                          "The alpha value for the zeta distribution returned by this RNG stream.",
                          DoubleValue(3.14),
                          MakeDoubleAccessor(&ZetaRandomVariable::m_alpha),
                          MakeDoubleChecker<double>());
    return tid;
}
 
ZetaRandomVariable::ZetaRandomVariable()
{
    // m_alpha is initialized after constructor by attributes
    NS_LOG_FUNCTION(this);
}
 
double
ZetaRandomVariable::GetAlpha() const
{
    NS_LOG_FUNCTION(this);
    return m_alpha;
}
 
double
ZetaRandomVariable::GetValue(double alpha)
{
    NS_LOG_FUNCTION(this << alpha);
    m_b = std::pow(2.0, alpha - 1.0);
 
    double u;
    double v;
    double X;
    double T;
    double test;
 
    do
    {
        // Get a uniform random variable in [0,1].
        u = Peek()->RandU01();
        if (IsAntithetic())
        {
            u = (1 - u);
        }
 
        // Get a uniform random variable in [0,1].
        v = Peek()->RandU01();
        if (IsAntithetic())
        {
            v = (1 - v);
        }
 
        X = std::floor(std::pow(u, -1.0 / (m_alpha - 1.0)));
        T = std::pow(1.0 + 1.0 / X, m_alpha - 1.0);
        test = v * X * (T - 1.0) / (m_b - 1.0);
    } while (test > (T / m_b));
 
    return X;
}
 
uint32_t
ZetaRandomVariable::GetInteger(uint32_t alpha)
{
    NS_LOG_FUNCTION(this << alpha);
    return static_cast<uint32_t>(GetValue(alpha));
}
 
double
ZetaRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    return GetValue(m_alpha);
}
 
NS_OBJECT_ENSURE_REGISTERED(DeterministicRandomVariable);
 
TypeId
DeterministicRandomVariable::GetTypeId()
{
    static TypeId tid = TypeId("ns3::DeterministicRandomVariable")
                            .SetParent<RandomVariableStream>()
                            .SetGroupName("Core")
                            .AddConstructor<DeterministicRandomVariable>();
    return tid;
}
 
DeterministicRandomVariable::DeterministicRandomVariable()
    : m_count(0),
      m_next(0),
      m_data(nullptr)
{
    NS_LOG_FUNCTION(this);
}
 
DeterministicRandomVariable::~DeterministicRandomVariable()
{
    // Delete any values currently set.
    NS_LOG_FUNCTION(this);
    if (m_data != nullptr)
    {
        delete[] m_data;
    }
}
 
void
DeterministicRandomVariable::SetValueArray(const std::vector<double>& values)
{
    SetValueArray(values.data(), values.size());
}
 
void
DeterministicRandomVariable::SetValueArray(const double* values, std::size_t length)
{
    NS_LOG_FUNCTION(this << values << length);
    // Delete any values currently set.
    if (m_data != nullptr)
    {
        delete[] m_data;
    }
 
    // Make room for the values being set.
    m_data = new double[length];
    m_count = length;
    m_next = length;
 
    // Copy the values.
    for (std::size_t i = 0; i < m_count; i++)
    {
        m_data[i] = values[i];
    }
}
 
double
DeterministicRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
    // Make sure the array has been set.
    NS_ASSERT(m_count > 0);
 
    if (m_next == m_count)
    {
        m_next = 0;
    }
    return m_data[m_next++];
}
 
NS_OBJECT_ENSURE_REGISTERED(EmpiricalRandomVariable);
 
// ValueCDF methods
EmpiricalRandomVariable::ValueCDF::ValueCDF()
    : value(0.0),
      cdf(0.0)
{
    NS_LOG_FUNCTION(this);
}
 
EmpiricalRandomVariable::ValueCDF::ValueCDF(double v, double c)
    : value(v),
      cdf(c)
{
    NS_LOG_FUNCTION(this << v << c);
    NS_ASSERT(c >= 0.0 && c <= 1.0);
}
 
bool
operator<(EmpiricalRandomVariable::ValueCDF a, EmpiricalRandomVariable::ValueCDF b)
{
    return a.cdf < b.cdf;
}
 
TypeId
EmpiricalRandomVariable::GetTypeId()
{
    static TypeId tid =
        TypeId("ns3::EmpiricalRandomVariable")
            .SetParent<RandomVariableStream>()
            .SetGroupName("Core")
            .AddConstructor<EmpiricalRandomVariable>()
            .AddAttribute("Interpolate",
                          "Treat the CDF as a smooth distribution and interpolate, "
                          "default is to treat the CDF as a histogram and sample.",
                          BooleanValue(false),
                          MakeBooleanAccessor(&EmpiricalRandomVariable::m_interpolate),
                          MakeBooleanChecker());
    return tid;
}
 
EmpiricalRandomVariable::EmpiricalRandomVariable()
    : m_validated(false)
{
    NS_LOG_FUNCTION(this);
}
 
bool
EmpiricalRandomVariable::SetInterpolate(bool interpolate)
{
    NS_LOG_FUNCTION(this << interpolate);
    bool prev = m_interpolate;
    m_interpolate = interpolate;
    return prev;
}
 
bool
EmpiricalRandomVariable::PreSample(double& value)
{
    NS_LOG_FUNCTION(this);
 
    if (!m_validated)
    {
        Validate();
    }
 
    // Get a uniform random variable in [0, 1].
    double r = Peek()->RandU01();
    if (IsAntithetic())
    {
        r = (1 - r);
    }
 
    value = r;
    bool valid = false;
    // check extrema
    if (r <= m_emp.front().cdf)
    {
        value = m_emp.front().value; // Less than first
        valid = true;
    }
    else if (r >= m_emp.back().cdf)
    {
        value = m_emp.back().value; // Greater than last
        valid = true;
    }
    return valid;
}
 
double
EmpiricalRandomVariable::GetValue()
{
    NS_LOG_FUNCTION(this);
 
    double value;
    if (PreSample(value))
    {
        return value;
    }
 
    // value now has the (unused) URNG selector
    if (m_interpolate)
    {
        value = DoInterpolate(value);
    }
    else
    {
        value = DoSampleCDF(value);
    }
    return value;
}
 
double
EmpiricalRandomVariable::DoSampleCDF(double r)
{
    NS_LOG_FUNCTION(this << r);
 
    ValueCDF selector(0, r);
    auto bound = std::upper_bound(m_emp.begin(), m_emp.end(), selector);
 
    return bound->value;
}
 
double
EmpiricalRandomVariable::Interpolate()
{
    NS_LOG_FUNCTION(this);
 
    double value;
    if (PreSample(value))
    {
        return value;
    }
 
    // value now has the (unused) URNG selector
    value = DoInterpolate(value);
    return value;
}
 
double
EmpiricalRandomVariable::DoInterpolate(double r)
{
    NS_LOG_FUNCTION(this << r);
 
    // Return a value from the empirical distribution
    // This code based (loosely) on code by Bruce Mah (Thanks Bruce!)
 
    // search
    ValueCDF selector(0, r);
    auto upper = std::upper_bound(m_emp.begin(), m_emp.end(), selector);
    auto lower = std::prev(upper, 1);
    if (upper == m_emp.begin())
    {
        lower = upper;
    }
 
    // Interpolate random value in range [v1..v2) based on [c1 .. r .. c2)
    double c1 = lower->cdf;
    double c2 = upper->cdf;
    double v1 = lower->value;
    double v2 = upper->value;
 
    double value = (v1 + ((v2 - v1) / (c2 - c1)) * (r - c1));
    return value;
}
 
void
EmpiricalRandomVariable::CDF(double v, double c)
{
    // Add a new empirical datapoint to the empirical cdf
    // NOTE.   These MUST be inserted in non-decreasing order
    NS_LOG_FUNCTION(this << v << c);
    m_emp.emplace_back(v, c);
}
 
void
EmpiricalRandomVariable::Validate()
{
    NS_LOG_FUNCTION(this);
    if (m_emp.empty())
    {
        NS_FATAL_ERROR("CDF is not initialized");
    }
    ValueCDF prior = m_emp[0];
    for (auto current : m_emp)
    {
        if (current.value < prior.value || current.cdf < prior.cdf)
        { // Error
            std::cerr << "Empirical Dist error,"
                      << " current value " << current.value << " prior value " << prior.value
                      << " current cdf " << current.cdf << " prior cdf " << prior.cdf << std::endl;
            NS_FATAL_ERROR("Empirical Dist error");
        }
        prior = current;
    }
    if (prior.cdf != 1.0)
    {
        NS_FATAL_ERROR("CDF does not cover the whole distribution");
    }
    m_validated = true;
}
 
} // namespace ns3