gtsam/tests/testSubgraphPreconditioner.cpp

221 lines
7.5 KiB
C++
Raw Permalink Normal View History

/* ----------------------------------------------------------------------------
2019-04-08 10:33:58 +08:00
* GTSAM Copyright 2010, Georgia Tech Research Corporation,
* Atlanta, Georgia 30332-0415
* All Rights Reserved
* Authors: Frank Dellaert, et al. (see THANKS for the full author list)
* See LICENSE for the license information
* -------------------------------------------------------------------------- */
/**
* @file testSubgraphConditioner.cpp
* @brief Unit tests for SubgraphPreconditioner
* @author Frank Dellaert
**/
#include <tests/smallExample.h>
2019-04-08 10:33:58 +08:00
#include <gtsam/base/numericalDerivative.h>
#include <gtsam/inference/Ordering.h>
#include <gtsam/inference/Symbol.h>
2019-04-05 11:01:47 +08:00
#include <gtsam/linear/GaussianEliminationTree.h>
2019-04-08 10:33:58 +08:00
#include <gtsam/linear/GaussianFactorGraph.h>
#include <gtsam/linear/SubgraphPreconditioner.h>
2019-04-08 10:33:58 +08:00
#include <gtsam/linear/iterative.h>
#include <gtsam/slam/dataset.h>
#include <gtsam/symbolic/SymbolicFactorGraph.h>
#include <CppUnitLite/TestHarness.h>
#include <fstream>
using namespace std;
using namespace gtsam;
using namespace example;
// define keys
// Create key for simulated planar graph
2019-04-08 10:33:58 +08:00
Symbol key(int x, int y) { return symbol_shorthand::X(1000 * x + y); }
/* ************************************************************************* */
2019-04-08 10:33:58 +08:00
TEST(SubgraphPreconditioner, planarOrdering) {
// Check canonical ordering
Ordering ordering = planarOrdering(3),
expected{key(3, 3), key(2, 3), key(1, 3), key(3, 2), key(2, 2),
key(1, 2), key(3, 1), key(2, 1), key(1, 1)};
2019-04-08 10:33:58 +08:00
EXPECT(assert_equal(expected, ordering));
}
/* ************************************************************************* */
/** unnormalized error */
2013-08-06 06:31:44 +08:00
static double error(const GaussianFactorGraph& fg, const VectorValues& x) {
double total_error = 0.;
2019-04-08 10:33:58 +08:00
for (const GaussianFactor::shared_ptr& factor : fg)
total_error += factor->error(x);
return total_error;
}
/* ************************************************************************* */
2019-04-08 10:33:58 +08:00
TEST(SubgraphPreconditioner, planarGraph) {
// Check planar graph construction
const auto [A, xtrue] = planarGraph(3);
2019-04-08 10:33:58 +08:00
LONGS_EQUAL(13, A.size());
LONGS_EQUAL(9, xtrue.size());
DOUBLES_EQUAL(0, error(A, xtrue), 1e-9); // check zero error for xtrue
// Check that xtrue is optimal
2021-12-31 01:18:49 +08:00
GaussianBayesNet R1 = *A.eliminateSequential();
VectorValues actual = R1.optimize();
2019-04-08 10:33:58 +08:00
EXPECT(assert_equal(xtrue, actual));
}
/* ************************************************************************* */
2019-04-08 10:33:58 +08:00
TEST(SubgraphPreconditioner, splitOffPlanarTree) {
// Build a planar graph
const auto [A, xtrue] = planarGraph(3);
// Get the spanning tree and constraints, and check their sizes
const auto [T, C] = splitOffPlanarTree(3, A);
LONGS_EQUAL(9, T.size());
LONGS_EQUAL(4, C.size());
// Check that the tree can be solved to give the ground xtrue
2021-12-31 01:18:49 +08:00
GaussianBayesNet R1 = *T.eliminateSequential();
VectorValues xbar = R1.optimize();
2019-04-08 10:33:58 +08:00
EXPECT(assert_equal(xtrue, xbar));
}
/* ************************************************************************* */
2019-04-08 10:33:58 +08:00
TEST(SubgraphPreconditioner, system) {
// Build a planar graph
size_t N = 3;
const auto [Ab, xtrue] = planarGraph(N); // A*x-b
2019-04-05 11:01:47 +08:00
// Get the spanning tree and remaining graph
auto [Ab1, Ab2] = splitOffPlanarTree(N, Ab);
// Eliminate the spanning tree to build a prior
2019-04-05 11:01:47 +08:00
const Ordering ord = planarOrdering(N);
2021-12-31 01:18:49 +08:00
auto Rc1 = *Ab1.eliminateSequential(ord); // R1*x-c1
VectorValues xbar = Rc1.optimize(); // xbar = inv(R1)*c1
// Create Subgraph-preconditioned system
const SubgraphPreconditioner system(Ab2, Rc1, xbar);
2019-04-05 11:01:47 +08:00
// Get corresponding matrices for tests. Add dummy factors to Ab2 to make
// sure it works with the ordering.
Ordering ordering = Rc1.ordering(); // not ord in general!
Ab2.add(key(1, 1), Z_2x2, Z_2x1);
Ab2.add(key(1, 2), Z_2x2, Z_2x1);
Ab2.add(key(1, 3), Z_2x2, Z_2x1);
const auto [A, b] = Ab.jacobian(ordering);
const auto [A1, b1] = Ab1.jacobian(ordering);
const auto [A2, b2] = Ab2.jacobian(ordering);
Matrix R1 = Rc1.matrix(ordering).first;
2019-04-05 11:01:47 +08:00
Matrix Abar(13 * 2, 9 * 2);
Abar.topRows(9 * 2) = Matrix::Identity(9 * 2, 9 * 2);
Abar.bottomRows(8) = A2.topRows(8) * R1.inverse();
2019-04-05 11:01:47 +08:00
// Helper function to vectorize in correct order, which is the order in which
// we eliminated the spanning tree.
2019-04-08 10:33:58 +08:00
auto vec = [ordering](const VectorValues& x) { return x.vector(ordering); };
// Set up y0 as all zeros
const VectorValues y0 = system.zero();
// y1 = perturbed y0
VectorValues y1 = system.zero();
2019-04-08 10:33:58 +08:00
y1[key(3, 3)] = Vector2(1.0, -1.0);
2019-04-05 11:01:47 +08:00
// Check backSubstituteTranspose works with R1
VectorValues actual = Rc1.backSubstituteTranspose(y1);
2019-04-05 11:01:47 +08:00
Vector expected = R1.transpose().inverse() * vec(y1);
EXPECT(assert_equal(expected, vec(actual)));
2019-04-05 11:01:47 +08:00
// Check corresponding x values
// for y = 0, we get xbar:
EXPECT(assert_equal(xbar, system.x(y0)));
// for non-zero y, answer is x = xbar + inv(R1)*y
const Vector expected_x1 = vec(xbar) + R1.inverse() * vec(y1);
const VectorValues x1 = system.x(y1);
EXPECT(assert_equal(expected_x1, vec(x1)));
// Check errors
2019-04-08 10:33:58 +08:00
DOUBLES_EQUAL(0, error(Ab, xbar), 1e-9);
DOUBLES_EQUAL(0, system.error(y0), 1e-9);
DOUBLES_EQUAL(2, error(Ab, x1), 1e-9);
DOUBLES_EQUAL(2, system.error(y1), 1e-9);
2019-04-05 11:01:47 +08:00
// Check that transposeMultiplyAdd <=> y += alpha * Abar' * e
// We check for e1 =[1;0] and e2=[0;1] corresponding to T and C
const double alpha = 0.5;
2019-04-08 10:33:58 +08:00
Errors e1, e2;
for (size_t i = 0; i < 13; i++) {
e1.push_back(i < 9 ? Vector2(1, 1) : Vector2(0, 0));
e2.push_back(i >= 9 ? Vector2(1, 1) : Vector2(0, 0));
2019-04-05 11:01:47 +08:00
}
2019-04-08 10:33:58 +08:00
Vector ee1(13 * 2), ee2(13 * 2);
ee1 << Vector::Ones(9 * 2), Vector::Zero(4 * 2);
ee2 << Vector::Zero(9 * 2), Vector::Ones(4 * 2);
2019-04-05 11:01:47 +08:00
// Check transposeMultiplyAdd for e1
VectorValues y = system.zero();
2019-04-05 11:01:47 +08:00
system.transposeMultiplyAdd(alpha, e1, y);
Vector expected_y = alpha * Abar.transpose() * ee1;
EXPECT(assert_equal(expected_y, vec(y)));
// Check transposeMultiplyAdd for e2
y = system.zero();
2019-04-05 11:01:47 +08:00
system.transposeMultiplyAdd(alpha, e2, y);
expected_y = alpha * Abar.transpose() * ee2;
EXPECT(assert_equal(expected_y, vec(y)));
// Test gradient in y
2019-04-05 11:01:47 +08:00
auto g = system.gradient(y0);
Vector expected_g = Vector::Zero(18);
EXPECT(assert_equal(expected_g, vec(g)));
}
/* ************************************************************************* */
2019-04-08 10:33:58 +08:00
TEST(SubgraphPreconditioner, conjugateGradients) {
// Build a planar graph
size_t N = 3;
const auto [Ab, xtrue] = planarGraph(N); // A*x-b
2019-04-05 11:01:47 +08:00
// Get the spanning tree
const auto [Ab1, Ab2] = splitOffPlanarTree(N, Ab);
// Eliminate the spanning tree to build a prior
2021-12-31 01:18:49 +08:00
GaussianBayesNet Rc1 = *Ab1.eliminateSequential(); // R1*x-c1
VectorValues xbar = Rc1.optimize(); // xbar = inv(R1)*c1
// Create Subgraph-preconditioned system
SubgraphPreconditioner system(Ab2, Rc1, xbar);
// Create zero config y0 and perturbed config y1
2013-08-06 06:31:44 +08:00
VectorValues y0 = VectorValues::Zero(xbar);
2013-08-06 06:31:44 +08:00
VectorValues y1 = y0;
2019-04-05 11:01:47 +08:00
y1[key(2, 2)] = Vector2(1.0, -1.0);
2013-08-06 06:31:44 +08:00
VectorValues x1 = system.x(y1);
// Solve for the remaining constraints using PCG
ConjugateGradientParameters parameters;
2013-08-06 06:31:44 +08:00
VectorValues actual = conjugateGradients<SubgraphPreconditioner,
VectorValues, Errors>(system, y1, parameters);
2019-04-05 11:01:47 +08:00
EXPECT(assert_equal(y0,actual));
// Compare with non preconditioned version:
2013-08-06 06:31:44 +08:00
VectorValues actual2 = conjugateGradientDescent(Ab, x1, parameters);
2019-04-08 10:33:58 +08:00
EXPECT(assert_equal(xtrue, actual2, 1e-4));
}
/* ************************************************************************* */
2019-04-08 10:33:58 +08:00
int main() {
TestResult tr;
return TestRegistry::runAllTests(tr);
}
/* ************************************************************************* */