def graph_implementation(arg_objs, size, data=None): """Reduces the atom to an affine expression and list of constraints. Parameters ---------- arg_objs : list LinExpr for each argument. size : tuple The size of the resulting expression. data : Additional data required by the atom. Returns ------- tuple (LinOp for objective, list of constraints) """ A = arg_objs[0] n, _ = A.size # SDP constraint. t = lu.create_var((1, 1)) prom_t = lu.promote(t, (n, 1)) # I*t - A expr = lu.sub_expr(lu.diag_vec(prom_t), A) return (t, [SDP(expr)])
def graph_implementation(arg_objs, size, data=None): """Reduces the atom to an affine expression and list of constraints. Parameters ---------- arg_objs : list LinExpr for each argument. size : tuple The size of the resulting expression. data : Additional data required by the atom. Returns ------- tuple (LinOp for objective, list of constraints) """ A = arg_objs[0] # n by m matrix. n, m = A.size # Create a matrix with Schur complement I*t - (1/t)*A.T*A. X = lu.create_var((n + m, n + m)) t = lu.create_var((1, 1)) constraints = [] # Fix X using the fact that A must be affine by the DCP rules. # X[0:n, 0:n] == I_n*t prom_t = lu.promote(t, (n, 1)) index.block_eq(X, lu.diag_vec(prom_t), constraints, 0, n, 0, n) # X[0:n, n:n+m] == A index.block_eq(X, A, constraints, 0, n, n, n + m) # X[n:n+m, n:n+m] == I_m*t prom_t = lu.promote(t, (m, 1)) # prom_t = lu.promote(lu.create_const(1, (1,1)), (m, 1)) index.block_eq(X, lu.diag_vec(prom_t), constraints, n, n + m, n, n + m) # Add SDP constraint. return (t, constraints + [SDP(X)])
def graph_implementation(arg_objs, size, data=None): """Reduces the atom to an affine expression and list of constraints. Parameters ---------- arg_objs : list LinExpr for each argument. size : tuple The size of the resulting expression. data : Additional data required by the atom. Returns ------- tuple (LinOp for objective, list of constraints) """ A = arg_objs[0] rows, cols = A.size # Create the equivalent problem: # minimize (trace(U) + trace(V))/2 # subject to: # [U A; A.T V] is positive semidefinite X = lu.create_var((rows + cols, rows + cols)) constraints = [] # Fix X using the fact that A must be affine by the DCP rules. # X[0:rows,rows:rows+cols] == A index.block_eq(X, A, constraints, 0, rows, rows, rows + cols) half = lu.create_const(0.5, (1, 1)) trace = lu.mul_expr(half, lu.trace(X), (1, 1)) # Add SDP constraint. return (trace, [SDP(X)] + constraints)
def graph_implementation(arg_objs, size, data=None): """Reduces the atom to an affine expression and list of constraints. Parameters ---------- arg_objs : list LinExpr for each argument. size : tuple The size of the resulting expression. data : Additional data required by the atom. Returns ------- tuple (LinOp for objective, list of constraints) """ X = arg_objs[0] # n by m matrix. P = arg_objs[1] # n by n matrix. n, m = X.size # Create a matrix with Schur complement T - X.T*P^-1*X. M = lu.create_var((n + m, n + m)) T = lu.create_var((m, m)) constraints = [] # Fix M using the fact that P must be affine by the DCP rules. # M[0:n, 0:n] == P. index.block_eq(M, P, constraints, 0, n, 0, n) # M[0:n, n:n+m] == X index.block_eq(M, X, constraints, 0, n, n, n + m) # M[n:n+m, n:n+m] == T index.block_eq(M, T, constraints, n, n + m, n, n + m) # Add SDP constraint. return (lu.trace(T), constraints + [SDP(M)])
def graph_implementation(arg_objs, size, data=None): """Reduces the atom to an affine expression and list of constraints. Parameters ---------- arg_objs : list LinExpr for each argument. size : tuple The size of the resulting expression. data : Additional data required by the atom. Returns ------- tuple (LinOp for objective, list of constraints) """ x = arg_objs[0] P = arg_objs[1] # n by n matrix. n, _ = P.size # Create a matrix with Schur complement t - x.T*P^-1*x. M = lu.create_var((n + 1, n + 1)) t = lu.create_var((1, 1)) constraints = [] # Fix M using the fact that P must be affine by the DCP rules. # M[0:n, 0:n] == P. index.block_eq(M, P, constraints, 0, n, 0, n) # M[0:n, n:n+1] == x index.block_eq(M, x, constraints, 0, n, n, n + 1) # M[n:n+1, n:n+1] == t index.block_eq(M, t, constraints, n, n + 1, n, n + 1) # Add SDP constraint. return (t, constraints + [SDP(M)])
def canonicalize(self): """Variable must be semidefinite and symmetric. """ upper_tri = lu.create_var((self.size[0], 1), self.id) fill_coeff = upper_tri_to_full(self.n) fill_coeff = lu.create_const(fill_coeff, (self.n*self.n, self.size[0]), sparse=True) full_mat = lu.mul_expr(fill_coeff, upper_tri, (self.n*self.n, 1)) full_mat = lu.reshape(full_mat, (self.n, self.n)) return (upper_tri, [SDP(full_mat, enforce_sym=False)])
def canonicalize(self): """Returns the graph implementation of the object. Marks the top level constraint as the dual_holder, so the dual value will be saved to the EqConstraint. Returns: A tuple of (affine expression, [constraints]). """ obj, constraints = self._expr.canonical_form half = lu.create_const(0.5, (1, 1)) symm = lu.mul_expr(half, lu.sum_expr([obj, lu.transpose(obj)]), obj.size) dual_holder = SDP(symm, enforce_sym=False, constr_id=self.id) return (None, constraints + [dual_holder])
def graph_implementation(arg_objs, size, data=None): """Reduces the atom to an affine expression and list of constraints. Creates the equivalent problem:: maximize sum(log(D[i, i])) subject to: D diagonal diag(D) = diag(Z) Z is upper triangular. [D Z; Z.T A] is positive semidefinite The problem computes the LDL factorization: .. math:: A = (Z^TD^{-1})D(D^{-1}Z) This follows from the inequality: .. math:: \det(A) >= \det(D) + \det([D, Z; Z^T, A])/\det(D) >= \det(D) because (Z^TD^{-1})D(D^{-1}Z) is a feasible D, Z that achieves det(A) = det(D) and the objective maximizes det(D). Parameters ---------- arg_objs : list LinExpr for each argument. size : tuple The size of the resulting expression. data : Additional data required by the atom. Returns ------- tuple (LinOp for objective, list of constraints) """ A = arg_objs[0] # n by n matrix. n, _ = A.size X = lu.create_var((2 * n, 2 * n)) X, constraints = Semidef(2 * n).canonical_form Z = lu.create_var((n, n)) D = lu.create_var((n, 1)) # Require that X and A are PSD. constraints += [SDP(A)] # Fix Z as upper triangular, D as diagonal, # and diag(D) as diag(Z). Z_lower_tri = lu.upper_tri(lu.transpose(Z)) constraints.append(lu.create_eq(Z_lower_tri)) # D[i, i] = Z[i, i] constraints.append(lu.create_eq(D, lu.diag_mat(Z))) # Fix X using the fact that A must be affine by the DCP rules. # X[0:n, 0:n] == D index.block_eq(X, lu.diag_vec(D), constraints, 0, n, 0, n) # X[0:n, n:2*n] == Z, index.block_eq(X, Z, constraints, 0, n, n, 2 * n) # X[n:2*n, n:2*n] == A index.block_eq(X, A, constraints, n, 2 * n, n, 2 * n) # Add the objective sum(log(D[i, i]) obj, constr = log.graph_implementation([D], (n, 1)) return (lu.sum_entries(obj), constraints + constr)