Hi Anders you probably should read my email on this of a few days ago. The simple answer is to use typedef boost::variate_generator<engine_t&, distribution_t> rng_t; (i.e. with ref) It all makes sense, it's just hard to get from the doc in my opinion. Kris

// The following program illustrates a problem in Boost.Random.

// Is this the desired behavior?

// If so, what is the reason for this?

// I use Boost version 1.31 and GCC version 3.3.3 on Suse 9.1

#include <vector> #include <algorithm> #include <boost/random.hpp>

// Definition of the random number generator typedef boost::mt19937 engine_t; typedef boost::uniform_int<int> distribution_t; typedef boost::variate_generator<engine_t, distribution_t> rng_t;

int main() { const unsigned int N = 13, A = 0, B = 1000; std::vector<int> run1(N), run2(N), run3(N);

// Generate some random integers { engine_t engine; distribution_t distribution(A,B); rng_t rng(engine, distribution); std::generate(run1.begin(), run1.end(), rng); }

// Verify that I get the same sequence if I start again { engine_t engine; distribution_t distribution(A,B); rng_t rng(engine, distribution); std::generate(run2.begin(), run2.end(), rng); }

// It appears that the generator restarts for each generate, since // the generate algorithm creates a copy of the functor so that // its state is lost { engine_t engine; distribution_t distribution(A,B); rng_t rng(engine, distribution); std::generate(run3.begin(), run3.begin()+N/2, rng); std::generate(run3.begin()+N/2, run3.end(), rng); }

// For me it would seem more natural if the engine class would be a // singleton, so that there can be only one object of any particular // random generator engine type.

// Our purpose for using random numbers is to simulate physical processes. // I can't think of a case when there would be a need for several random // engines, even if there can of course be many different random // distributions in an application.

// Having the same random number sequence in different parts of a simulation // can probably create very strange behaviour, since the numbers from the // different engines of the same type will be perfectly correlated.

// Perhaps other applications have use of several random engines?

for(unsigned i=0; i<N; ++i) { if(i==N/2) std::cout<<"Halfway there!"<<std::endl; std::cout <<run1[i] <<" == "<<run2[i] <<" == "<<run3[i] <<std::endl; }

return 0; }

// Anders Edin // Sidec Technologies AB