Skip to content

Commit

Permalink
Activate all examples
Browse files Browse the repository at this point in the history
  • Loading branch information
@vfisikop
vfisikop committed Jul 4, 2024
1 parent dce66d6 commit 9be4f78
Showing 1 changed file with 39 additions and 54 deletions.
93 changes: 39 additions & 54 deletions examples/sample_points/simple_example.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -130,7 +130,7 @@ template< typename SpMat, typename VT>
void load_crhmc_problem(SpMat &A, VT &b, VT &lb, VT &ub, int &dimension,
std::string problem_name) {
{
std::string fileName("../crhmc_sampling/data/");
std::string fileName("./data/");
fileName.append(problem_name);
fileName.append(".mm");
if(!exists_check(fileName)){
Expand All @@ -144,7 +144,7 @@ void load_crhmc_problem(SpMat &A, VT &b, VT &lb, VT &ub, int &dimension,
b = VT(X.col(dimension));
}
{
std::string fileName("../crhmc_sampling/data/");
std::string fileName("./data/");
fileName.append(problem_name);
fileName.append("_bounds.mm");
if(!exists_check(fileName)){
Expand All @@ -159,12 +159,10 @@ void load_crhmc_problem(SpMat &A, VT &b, VT &lb, VT &ub, int &dimension,


int main() {
// NEW INTERFACE Sampling

// Inputs:
// 1. Inputs:

int d = 100;
int rnum = 1000;
int d = 10;

// Generating a 3-dimensional cube centered at origin
HPolytopeType HP = generate_cube<HPolytopeType>(d, false);
Expand All @@ -176,7 +174,7 @@ int main() {
// Setup parameters for sampling
Point q(HP.dimension());
RNGType rng(HP.dimension());
/*

// Generating a sparse polytope/problem
using SpMat = Eigen::SparseMatrix<NT>;
using ConstraintProblem =constraint_problem<SpMat, Point>;
Expand All @@ -189,9 +187,10 @@ int main() {
ConstraintProblem problem = ConstraintProblem(dimension);
problem.set_equality_constraints(As, b);
problem.set_bounds(lb, ub);
*/

// Walks

// 2. Walks

AcceleratedBilliardWalk abill_walk;
AcceleratedBilliardWalk abill_walk_custom(10); //user defined walk parameters
BallWalk ball_walk;
Expand All @@ -216,12 +215,12 @@ int main() {
UnderdampedLangevinWalk uld_walk;
CRHMCWalk crhmc_walk;

// Distributions
// 3. Distributions

// 1. Uniform
// Uniform
UniformDistribution udistr{};

// 2. Spherical
// Spherical Gaussian
SphericalGaussianDistribution sgdistr{};

MT A(2, 2);
Expand All @@ -230,12 +229,12 @@ int main() {
Ellipsoid<Point> ell(A); // origin centered ellipsoid
GaussianDistribution gdistr(ell);

// 3. Exponential
// Exponential
NT variance = 1.0;
auto c = GetDirection<Point>::apply(HP.dimension(), rng, false);
ExponentialDistribution edistr(c, variance);

// 4. LogConcave
// LogConcave

std::pair<Point, NT> inner_ball = HP.ComputeInnerBall();
Point x0 = inner_ball.first;
Expand Down Expand Up @@ -269,13 +268,13 @@ int main() {



// Sampling
// 4. Sampling

using NT = double;
using MT = Eigen::Matrix<NT,Eigen::Dynamic,Eigen::Dynamic>;
using VT = Eigen::Matrix<NT,Eigen::Dynamic,1>;

//int rnum = 110;
int rnum = 100;
int nburns = 5;
int walk_len = 1;

Expand All @@ -284,52 +283,38 @@ int main() {
// 1. the eigen matrix interface
std::cout << "uniform" << std::endl;
sample_points_eigen_matrix(HP, q, abill_walk, udistr, rng, 1, rnum, nburns);
//sample_points_eigen_matrix(HP, q, abill_walk_custom, udistr, rng, walk_len, rnum, nburns);
//sample_points_eigen_matrix(HP, q, ball_walk, udistr, rng, 2*d, rnum, nburns);
//sample_points_eigen_matrix(HP, q, cdhr_walk, udistr, rng, 2*d, rnum, nburns);
//sample_points_eigen_matrix(HP, q, dikin_walk, udistr, rng, walk_len, rnum, nburns);
//sample_points_eigen_matrix(HP, q, john_walk, udistr, rng, walk_len, rnum, nburns);
//sample_points_eigen_matrix(HP, q, rdhr_walk, udistr, rng, 2*d, rnum, nburns);
//sample_points_eigen_matrix(HP, q, vaidya_walk, udistr, rng, walk_len, rnum, nburns);

//Compute chebychev ball
std::pair<Point,NT> CheBall;
//CheBall = HP.ComputeInnerBall();

// accelerated billiard
auto start = std::chrono::steady_clock::now();
samples = accelerated_billiard_walk(HP, rng, 1, rnum);
auto end = std::chrono::steady_clock::now();
auto time = std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count();
std::cout << " time= " << time;
unsigned int min_ess;
auto score = effective_sample_size<NT, VT, MT>(samples, min_ess);
std::cout << "\t ess= " << min_ess << std::endl;
sample_points_eigen_matrix(HP, q, abill_walk_custom, udistr, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, ball_walk, udistr, rng, 2*d, rnum, nburns);
sample_points_eigen_matrix(HP, q, cdhr_walk, udistr, rng, 2*d, rnum, nburns);
sample_points_eigen_matrix(HP, q, dikin_walk, udistr, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, john_walk, udistr, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, rdhr_walk, udistr, rng, 2*d, rnum, nburns);
sample_points_eigen_matrix(HP, q, vaidya_walk, udistr, rng, walk_len, rnum, nburns);


// std::cout << "shperical gaussian" << std::endl;
// sample_points_eigen_matrix(HP, q, gball_walk, sgdistr, rng, walk_len, rnum, nburns);
// sample_points_eigen_matrix(HP, q, gcdhr_walk, sgdistr, rng, walk_len, rnum, nburns);
// sample_points_eigen_matrix(HP, q, grdhr_walk, sgdistr, rng, walk_len, rnum, nburns);
// sample_points_eigen_matrix(HP, q, ghmc_walk, sgdistr, rng, walk_len, rnum, nburns);
std::cout << "shperical gaussian" << std::endl;
sample_points_eigen_matrix(HP, q, gball_walk, sgdistr, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, gcdhr_walk, sgdistr, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, grdhr_walk, sgdistr, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, ghmc_walk, sgdistr, rng, walk_len, rnum, nburns);

// std::cout << "general gaussian" << std::endl;
// sample_points_eigen_matrix(HP, q, gbill_walk, gdistr, rng, walk_len, rnum, nburns);
std::cout << "general gaussian" << std::endl;
sample_points_eigen_matrix(HP, q, gbill_walk, gdistr, rng, walk_len, rnum, nburns);

// std::cout << "exponential" << std::endl;
// sample_points_eigen_matrix(HP, q, ehmc_walk, edistr, rng, walk_len, rnum, nburns);
std::cout << "exponential" << std::endl;
sample_points_eigen_matrix(HP, q, ehmc_walk, edistr, rng, walk_len, rnum, nburns);

std::cout << "logconcave" << std::endl;
// sample_points_eigen_matrix(HP, q, hmc_walk, logconcave_reflective, rng, walk_len, rnum, nburns);
// sample_points_eigen_matrix(HP, q, nhmc_walk, logconcave_reflective, rng, walk_len, rnum, nburns);
// sample_points_eigen_matrix(HP, q, nhmc_walk, logconcave_ref_gaus, rng, walk_len, rnum, nburns);
// sample_points_eigen_matrix(HP, q, uld_walk, logconcave_ref_gaus, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, hmc_walk, logconcave_reflective, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, nhmc_walk, logconcave_reflective, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, nhmc_walk, logconcave_ref_gaus, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, uld_walk, logconcave_ref_gaus, rng, walk_len, rnum, nburns);

// sample_points_eigen_matrix(HP, q, crhmc_walk, logconcave_crhmc, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, crhmc_walk, logconcave_crhmc, rng, walk_len, rnum, nburns);
// The following will compile but segfauls since walk and distribution are not compatible
//sample_points_eigen_matrix(HP, q, nhmc_walk, logconcave_crhmc, rng, walk_len, rnum, nburns);
// sample_points_eigen_matrix(problem, q, crhmc_walk, logconcave_crhmc, rng, walk_len, rnum, nburns);

//sample_points_eigen_matrix(problem, q, crhmc_walk, logconcave_uniform, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(problem, q, crhmc_walk, logconcave_crhmc, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(problem, q, crhmc_walk, logconcave_uniform, rng, walk_len, rnum, nburns);
sample_points_eigen_matrix(HP, q, crhmc_walk, logconcave_uniform, rng, walk_len, rnum, nburns);
//sample_points_eigen_matrix(problem, q, nhmc_walk, logconcave_uniform2, rng, walk_len, rnum, nburns);

Expand Down

0 comments on commit 9be4f78

Please sign in to comment.