1 // Ceres Solver - A fast non-linear least squares minimizer
2 // Copyright 2010, 2011, 2012 Google Inc. All rights reserved.
3 // http://code.google.com/p/ceres-solver/
4 //
5 // Redistribution and use in source and binary forms, with or without
6 // modification, are permitted provided that the following conditions are met:
7 //
8 // * Redistributions of source code must retain the above copyright notice,
9 // this list of conditions and the following disclaimer.
10 // * Redistributions in binary form must reproduce the above copyright notice,
11 // this list of conditions and the following disclaimer in the documentation
12 // and/or other materials provided with the distribution.
13 // * Neither the name of Google Inc. nor the names of its contributors may be
14 // used to endorse or promote products derived from this software without
15 // specific prior written permission.
16 //
17 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
18 // AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 // IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 // ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
21 // LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
22 // CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
23 // SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
24 // INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
25 // CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
26 // ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
27 // POSSIBILITY OF SUCH DAMAGE.
28 //
29 // Author: keir@google.com (Keir Mierle)
30 //
31 // Minimize 0.5 (10 - x)^2 using jacobian matrix computed using
32 // numeric differentiation.
33
34 #include <vector>
35 #include "ceres/ceres.h"
36 #include "gflags/gflags.h"
37 #include "glog/logging.h"
38
39 using ceres::NumericDiffCostFunction;
40 using ceres::CENTRAL;
41 using ceres::SizedCostFunction;
42 using ceres::CostFunction;
43 using ceres::Problem;
44 using ceres::Solver;
45 using ceres::Solve;
46
47 class ResidualWithNoDerivative
48 : public SizedCostFunction<1 /* number of residuals */,
49 1 /* size of first parameter */> {
50 public:
~ResidualWithNoDerivative()51 virtual ~ResidualWithNoDerivative() {}
Evaluate(double const * const * parameters,double * residuals,double ** jacobians) const52 virtual bool Evaluate(double const* const* parameters,
53 double* residuals,
54 double** jacobians) const {
55 (void) jacobians; // Ignored; filled in by numeric differentiation.
56
57 // f(x) = 10 - x.
58 residuals[0] = 10 - parameters[0][0];
59 return true;
60 }
61 };
62
main(int argc,char ** argv)63 int main(int argc, char** argv) {
64 google::ParseCommandLineFlags(&argc, &argv, true);
65 google::InitGoogleLogging(argv[0]);
66
67 // The variable to solve for with its initial value.
68 double initial_x = 5.0;
69 double x = initial_x;
70
71 // Set up the only cost function (also known as residual). This uses
72 // numeric differentiation to obtain the derivative (jacobian).
73 CostFunction* cost =
74 new NumericDiffCostFunction<ResidualWithNoDerivative, CENTRAL, 1, 1> (
75 new ResidualWithNoDerivative, ceres::TAKE_OWNERSHIP);
76
77 // Build the problem.
78 Problem problem;
79 problem.AddResidualBlock(cost, NULL, &x);
80
81 // Run the solver!
82 Solver::Options options;
83 options.max_num_iterations = 10;
84 options.linear_solver_type = ceres::DENSE_QR;
85 options.minimizer_progress_to_stdout = true;
86 Solver::Summary summary;
87 Solve(options, &problem, &summary);
88 std::cout << summary.BriefReport() << "\n";
89 std::cout << "x : " << initial_x
90 << " -> " << x << "\n";
91 return 0;
92 }
93