Is there a quadratic programming library in C++? [closed]

Question

The only Google search result I found is QuadProg++ but it can not solve the quadratic programming problem whose matrix is not applicable for Cholesky decomposition.

So can anyone give me some suggestion on other library? Thanks.

Answer 1

CGAL looks great for quadratic programming. There is even a manual.

  // by default, we have a nonnegative QP with Ax <= b
  Program qp (CGAL::SMALLER, true, 0, false, 0); 

  // now set the non-default entries: 
  const int X = 0; 
  const int Y = 1;
  qp.set_a(X, 0,  1); qp.set_a(Y, 0, 1); qp.set_b(0, 7);  //  x + y  <= 7
  qp.set_a(X, 1, -1); qp.set_a(Y, 1, 2); qp.set_b(1, 4);  // -x + 2y <= 4
  qp.set_u(Y, true, 4);                                   //       y <= 4
  qp.set_d(X, X, 2); qp.set_d (Y, Y, 8); // !!specify 2D!!    x^2 + 4 y^2
  qp.set_c(Y, -32);                                       // -32y
  qp.set_c0(64);                                          // +64

  // solve the program, using ET as the exact type
  Solution s = CGAL::solve_quadratic_program(qp, ET());
  assert (s.solves_quadratic_program(qp));

Code from the first example.

Answer 2

LAPACK has a number of Cholesky decomposition routines (they call it Cholesky factorization). There are C++ wrappers available for LAPACK (see this SO question for a list).

The answer from Anycom in that post is a bit cryptic, but he meant that there are LAPACK bindings that can be used together with Boost's linear algebra library, uBLAS.

---

I've found this library: OOQP (Object-Oriented Software for Quadratic Programming). If you scroll down that page, you'll find a research paper and a user guide. There seems to be a C++ API for that library. Hope this helps.