RhodeCode

      static Result convert(RandomCore<Word>& rnd) {

        return Shifting<Result, - shift - bits>::

          shift(static_cast<Result>(rnd())) +

          RealConversion<Result, Word, rest-bits, shift + bits>::

          convert(rnd);

      }

    };

    template <typename Result, typename Word>

    struct Initializer {

      template <typename Iterator>

      static void init(RandomCore<Word>& rnd, Iterator begin, Iterator end) {

        std::vector<Word> ws;

        for (Iterator it = begin; it != end; ++it) {

          ws.push_back(Word(*it));

        }

        rnd.initState(ws.begin(), ws.end());

      }

      static void init(RandomCore<Word>& rnd, Result seed) {

        rnd.initState(seed);

      }

    };

    template <typename Word>

    struct BoolConversion {

      static bool convert(RandomCore<Word>& rnd) {

        return (rnd() & 1) == 1;

      }

    };

    template <typename Word>

    struct BoolProducer {

      Word buffer;

      int num;

      BoolProducer() : num(0) {}

      bool convert(RandomCore<Word>& rnd) {

        if (num == 0) {

          buffer = rnd();

          num = RandomTraits<Word>::bits;

        }

        bool r = (buffer & 1);

        buffer >>= 1;

        --num;

        return r;

      }

    };

  }

  /// \ingroup misc

  ///

  /// \brief Mersenne Twister random number generator

  ///

  /// The Mersenne Twister is a twisted generalized feedback

  /// shift-register generator of Matsumoto and Nishimura. The period

  /// of this generator is \f$ 2^{19937} - 1 \f$ and it is

  /// equi-distributed in 623 dimensions for 32-bit numbers. The time

  /// performance of this generator is comparable to the commonly used

  /// generators.

  ///

  /// This implementation is specialized for both 32-bit and 64-bit

  /// architectures. The generators differ sligthly in the

  /// initialization and generation phase so they produce two

  /// completly different sequences.

  ///

  /// The generator gives back random numbers of serveral types. To

  /// get a random number from a range of a floating point type you

  /// can use one form of the \c operator() or the \c real() member

  /// function. If you want to get random number from the {0, 1, ...,

  /// n-1} integer range use the \c operator[] or the \c integer()

  /// method. And to get random number from the whole range of an

  /// integer type you can use the argumentless \c integer() or \c

  /// uinteger() functions. After all you can get random bool with

  /// equal chance of true and false or given probability of true

  /// result with the \c boolean() member functions.

  ///

  ///\code

  /// // The commented code is identical to the other

  /// double a = rnd();                     // [0.0, 1.0)

  /// // double a = rnd.real();             // [0.0, 1.0)

  /// double b = rnd(100.0);                // [0.0, 100.0)

  /// // double b = rnd.real(100.0);        // [0.0, 100.0)

  /// double c = rnd(1.0, 2.0);             // [1.0, 2.0)

  /// // double c = rnd.real(1.0, 2.0);     // [1.0, 2.0)

  /// int d = rnd[100000];                  // 0..99999

  /// // int d = rnd.integer(100000);       // 0..99999

  /// int e = rnd[6] + 1;                   // 1..6

  /// // int e = rnd.integer(1, 1 + 6);     // 1..6

  /// int b = rnd.uinteger<int>();          // 0 .. 2^31 - 1

  /// int c = rnd.integer<int>();           // - 2^31 .. 2^31 - 1

  /// bool g = rnd.boolean();               // P(g = true) = 0.5

  /// bool h = rnd.boolean(0.8);            // P(h = true) = 0.8

  ///\endcode

  ///

  /// LEMON provides a global instance of the random number

  /// generator which name is \ref lemon::rnd "rnd". Usually it is a

  /// good programming convenience to use this global generator to get

  /// random numbers.

  class Random {

  private:

    // Architecture word

    typedef unsigned long Word;

    _random_bits::RandomCore<Word> core;

    _random_bits::BoolProducer<Word> bool_producer;

  public:

    /// \brief Default constructor

    ///

    /// Constructor with constant seeding.

    Random() { core.initState(); }

    /// \brief Constructor with seed

    ///

    /// Constructor with seed. The current number type will be converted

    /// to the architecture word type.

    template <typename Number>

    Random(Number seed) { 

      _random_bits::Initializer<Number, Word>::init(core, seed);

    }

    /// \brief Constructor with array seeding

    ///

    /// Constructor with array seeding. The given range should contain

    /// any number type and the numbers will be converted to the

    /// architecture word type.

    template <typename Iterator>

    Random(Iterator begin, Iterator end) { 

      typedef typename std::iterator_traits<Iterator>::value_type Number;

      _random_bits::Initializer<Number, Word>::init(core, begin, end);

    }

    /// \brief Copy constructor

    ///

    /// Copy constructor. The generated sequence will be identical to

    /// the other sequence. It can be used to save the current state

    /// of the generator and later use it to generate the same

    /// sequence.

    Random(const Random& other) {

      core.copyState(other.core);

    }

    /// \brief Assign operator

    ///

    /// Assign operator. The generated sequence will be identical to

    /// the other sequence. It can be used to save the current state

    /// of the generator and later use it to generate the same

    /// sequence.

    Random& operator=(const Random& other) {

      if (&other != this) {

        core.copyState(other.core);

      }

      return *this;

    }

    /// \brief Seeding random sequence

    ///

    /// Seeding the random sequence. The current number type will be

    /// converted to the architecture word type.

    template <typename Number>

    void seed(Number seed) { 

      _random_bits::Initializer<Number, Word>::init(core, seed);

    }

    /// \brief Seeding random sequence

    ///

    /// Seeding the random sequence. The given range should contain

    /// any number type and the numbers will be converted to the

    /// architecture word type.

    template <typename Iterator>

    void seed(Iterator begin, Iterator end) { 

      typedef typename std::iterator_traits<Iterator>::value_type Number;

      _random_bits::Initializer<Number, Word>::init(core, begin, end);

    }

    /// \brief Returns a random real number from the range [0, 1)

    ///

    /// It returns a random real number from the range [0, 1). The

    /// default Number type is \c double.

    template <typename Number>

    Number real() {

      return _random_bits::RealConversion<Number, Word>::convert(core);

    }

    double real() {

      return real<double>();

    }

    /// \brief Returns a random real number the range [0, b)

    ///

    /// It returns a random real number from the range [0, b).

    template <typename Number>

    Number real(Number b) { 

      return real<Number>() * b; 

    }

    /// \brief Returns a random real number from the range [a, b)

    ///

    /// It returns a random real number from the range [a, b).

    template <typename Number>

    Number real(Number a, Number b) { 

      return real<Number>() * (b - a) + a; 

    }

    /// \brief Returns a random real number from the range [0, 1)

    ///

    /// It returns a random double from the range [0, 1).

    double operator()() {

      return real<double>();

    }

    /// \brief Returns a random real number from the range [0, b)

    ///

    /// It returns a random real number from the range [0, b).

    template <typename Number>

    Number operator()(Number b) { 

      return real<Number>() * b; 

    }

    /// \brief Returns a random real number from the range [a, b)

    ///

    /// It returns a random real number from the range [a, b).

    template <typename Number>

    Number operator()(Number a, Number b) { 

      return real<Number>() * (b - a) + a; 

    }

    /// \brief Returns a random integer from a range

    ///

    /// It returns a random integer from the range {0, 1, ..., b - 1}.

    template <typename Number>

    Number integer(Number b) {

      return _random_bits::Mapping<Number, Word>::map(core, b);

    }

    /// \brief Returns a random integer from a range

    ///

    /// It returns a random integer from the range {a, a + 1, ..., b - 1}.

    template <typename Number>

    Number integer(Number a, Number b) {

      return _random_bits::Mapping<Number, Word>::map(core, b - a) + a;

    }

    /// \brief Returns a random integer from a range

    ///

    /// It returns a random integer from the range {0, 1, ..., b - 1}.

    template <typename Number>

    Number operator[](Number b) {

      return _random_bits::Mapping<Number, Word>::map(core, b);

    }

    /// \brief Returns a random non-negative integer

    ///

    /// It returns a random non-negative integer uniformly from the

    /// whole range of the current \c Number type. The default result

    /// type of this function is <tt>unsigned int</tt>.

    template <typename Number>

    Number uinteger() {

      return _random_bits::IntConversion<Number, Word>::convert(core);

    }

    unsigned int uinteger() {

      return uinteger<unsigned int>();

    }

    /// \brief Returns a random integer

    ///

    /// It returns a random integer uniformly from the whole range of

    /// the current \c Number type. The default result type of this

    /// function is \c int.

    template <typename Number>

    Number integer() {

      static const int nb = std::numeric_limits<Number>::digits + 

        (std::numeric_limits<Number>::is_signed ? 1 : 0);

      return _random_bits::IntConversion<Number, Word, nb>::convert(core);

    }

    int integer() {

      return integer<int>();

    }

    /// \brief Returns a random bool

    ///

    /// It returns a random bool. The generator holds a buffer for

    /// random bits. Every time when it become empty the generator makes

    /// a new random word and fill the buffer up.

    bool boolean() {

      return bool_producer.convert(core);

    }

    ///\name Non-uniform distributions

    ///

    ///@{

    /// \brief Returns a random bool

    ///

    /// It returns a random bool with given probability of true result.

    bool boolean(double p) {

      return operator()() < p;

    }

    /// Standard Gauss distribution

    /// Standard Gauss distribution.

    /// \note The Cartesian form of the Box-Muller

    /// transformation is used to generate a random normal distribution.

    /// \todo Consider using the "ziggurat" method instead.

    double gauss() 

    {

      double V1,V2,S;

      do {

	V1=2*real<double>()-1;

	V2=2*real<double>()-1;

	S=V1*V1+V2*V2;

      } while(S>=1);

      return std::sqrt(-2*std::log(S)/S)*V1;

    }

    /// Gauss distribution with given mean and standard deviation

    /// Gauss distribution with given mean and standard deviation.

    /// \sa gauss()

    double gauss(double mean,double std_dev)

    {

      return gauss()*std_dev+mean;

    }

    /// Exponential distribution with given mean

    /// This function generates an exponential distribution random number

    /// with mean <tt>1/lambda</tt>.

    ///

    double exponential(double lambda=1.0)

    {

      return -std::log(1.0-real<double>())/lambda;

    }

    /// Gamma distribution with given integer shape

    /// This function generates a gamma distribution random number.

    /// 

    ///\param k shape parameter (<tt>k>0</tt> integer)

    double gamma(int k) 

    {

      double s = 0;

      for(int i=0;i<k;i++) s-=std::log(1.0-real<double>());

      return s;

    }

    /// Gamma distribution with given shape and scale parameter

    /// This function generates a gamma distribution random number.

    /// 

    ///\param k shape parameter (<tt>k>0</tt>)

    ///\param theta scale parameter

    ///

    double gamma(double k,double theta=1.0)

    {

      double xi,nu;

      const double delta = k-std::floor(k);

      const double v0=E/(E-delta);

      do {

	double V0=1.0-real<double>();

	double V1=1.0-real<double>();

	double V2=1.0-real<double>();

	if(V2<=v0) 

	  {

	    xi=std::pow(V1,1.0/delta);

	    nu=V0*std::pow(xi,delta-1.0);

	  }

	else 

	  {

	    xi=1.0-std::log(V1);

	    nu=V0*std::exp(-xi);

	  }

      } while(nu>std::pow(xi,delta-1.0)*std::exp(-xi));

    }

    /// Weibull distribution

    /// This function generates a Weibull distribution random number.

    /// 

    ///\param k shape parameter (<tt>k>0</tt>)

    ///\param lambda scale parameter (<tt>lambda>0</tt>)

    ///

    double weibull(double k,double lambda)

    {

      return lambda*pow(-std::log(1.0-real<double>()),1.0/k);

    }  

    /// Pareto distribution

    /// This function generates a Pareto distribution random number.

    /// 

    ///\param k shape parameter (<tt>k>0</tt>)

    ///\param x_min location parameter (<tt>x_min>0</tt>)

    ///

    double pareto(double k,double x_min)

    {

    }  

    /// Poisson distribution

    /// This function generates a Poisson distribution random number with

    /// parameter \c lambda.

    /// 

    /// The probability mass function of this distribusion is

    /// \f[ \frac{e^{-\lambda}\lambda^k}{k!} \f]

    /// \note The algorithm is taken from the book of Donald E. Knuth titled

    /// ''Seminumerical Algorithms'' (1969). Its running time is linear in the

    /// return value.

    int poisson(double lambda)

    {

      const double l = std::exp(-lambda);

      int k=0;

      double p = 1.0;

      do {

	k++;

	p*=real<double>();

      } while (p>=l);

      return k-1;

    }  

    ///@}

    ///\name Two dimensional distributions

    ///

    ///@{

    /// Uniform distribution on the full unit circle

    /// Uniform distribution on the full unit circle.

    ///

    dim2::Point<double> disc() 

    {

      double V1,V2;

      do {

	V1=2*real<double>()-1;

	V2=2*real<double>()-1;

      } while(V1*V1+V2*V2>=1);

      return dim2::Point<double>(V1,V2);

    }

    /// A kind of two dimensional Gauss distribution

    /// This function provides a turning symmetric two-dimensional distribution.

    /// Both coordinates are of standard normal distribution, but they are not

    /// independent.

    ///

    /// \note The coordinates are the two random variables provided by

    /// the Box-Muller method.

    dim2::Point<double> gauss2()

    {

      double V1,V2,S;

      do {

	V1=2*real<double>()-1;

	V2=2*real<double>()-1;

	S=V1*V1+V2*V2;

      } while(S>=1);

      double W=std::sqrt(-2*std::log(S)/S);

      return dim2::Point<double>(W*V1,W*V2);

    }

    /// A kind of two dimensional exponential distribution

    /// This function provides a turning symmetric two-dimensional distribution.

    /// The x-coordinate is of conditionally exponential distribution

    /// with the condition that x is positive and y=0. If x is negative and 

    /// y=0 then, -x is of exponential distribution. The same is true for the

    /// y-coordinate.

    dim2::Point<double> exponential2() 

    {

      double V1,V2,S;

      do {

	V1=2*real<double>()-1;

	V2=2*real<double>()-1;

	S=V1*V1+V2*V2;

      } while(S>=1);

      double W=-std::log(S)/S;

      return dim2::Point<double>(W*V1,W*V2);

    }

    ///@}    

  };

  extern Random rnd;

}

#endif