Summary of implemented functions

Operators to manipulate PICOS Expressions

Operator Interpretation
+ addition
+= inplace addition
- substraction
* multiplication
^ Hadamard (elementwise) product
| scalar product
/ division
** exponentiation
abs() Euclidean (or Frobenius) norm
[] slicing
& horizontal concatenation
// vertical concatenation
.T transposition
.H Hermitian transposition
.Tx partial transposition
.conj complex conjugate
.real real part
.imag imaginary part

Operators that create constraints

Operator Interpretation
< or <= less or equal
> or >= larger or equal
== equal
<< Löwner ordering \preceq, or set membership \in
>> Löwner ordering \succeq, or set membership \ni

functions that create affine expressions

function short doc
sum() sums a list of affine expressions
diag() diagonal matrix defined by its diagonal
diag_vect() vector of diagonal elements of a matrix
new_param() constant affine expression
trace() trace of a square affine expression
lowtri() vector of lower triangular elements
partial_transpose() partial transposition
partial_trace() partial trace

functions to create convex constraints

function short doc
geomean() geometric mean
norm() (generalized) L_p- norm
tracepow() trace of a *p*th matrix power
detrootn() *n*th root of determinant
sum_k_largest() sum of k largest elements
sum_k_smallest() sum of k smallest elements
sum_k_largest_lambda() sum of k largest eigenvalues
sum_k_smallest_lambda() sum of k smallest eigenvalues
lambda_max() largest eigenvalue
lambda_min() smallest eigenvalue

functions that create sets

function short doc
ball(r,p) a L_p- ball of radius r
simplex(a) a standard simplex \{x\geq 0: \Vert x \Vert_1 \leq a \}
a set of the form

\{ 0\leq x\leq 1: \Vert x \Vert_1 \leq a\}, or \{x: \Vert x \Vert_\infty \leq 1; \Vert x \Vert_1 \leq a\}

Other useful functions

To transform a problem

function short doc
convert_quad_to_socp() replaces quadratic constraints by equivalent second order cone constraints
to_real() transform complex SDP to equivalent real-valued SDP
dualize() returns Lagrangian dual of a problem

Get information on a problem

function short doc
get_variable(name) gets the variable object name
get_valued_variable(name) gets the value of the variable name
check_current_value_feasibility() are the current variable value feasible?
obj_value() objective for the current variable values
.type returns problem’s type

Other tools

function short doc
available_solvers() lists installed solvers
import_cbf() imports data from a .cbf file
eval_dict() evaluates a dictionary of picos variables (after a problem has been solved)
write_to_file() writes problem to a file