Processing math: 100%
SuperSCS  1.3.2
All Data Structures Files Functions Variables Typedefs Enumerations Enumerator Macros Pages
Data Fields
scs_work Struct Reference

Workspace for SCS. More...

#include <scs.h>

Data Fields

scs_int m
 Row dimension of A. More...
 
scs_int n
 Column dimension of A. More...
 
scs_int l
 Dimension of u_k, that is l = m + n + 1 . More...
 
scs_float *RESTRICT u
 Vector u_k=(x_k,y_k,\tau_k). More...
 
scs_float *RESTRICT v
 Vector v_k = Qu_k (used only in SCS, not in SuperSCS). More...
 
scs_float *RESTRICT u_t
 Vector \tilde{u}_k. More...
 
scs_float *RESTRICT u_prev
 Vector u_{k-1} of the previous iteration. More...
 
scs_float *RESTRICT u_b
 Vector \bar{u}_k. More...
 
scs_float *RESTRICT h
 Auxiliary variable used in projections on linear spaces. More...
 
scs_float *RESTRICT g
 Auxiliary variable of dimension l-1 used in projections on linear spaces. More...
 
scs_float *RESTRICT pr
 Primal residual vector. More...
 
scs_float *RESTRICT dr
 Dual residual vector. More...
 
scs_float *RESTRICT b
 The (possibly normalized) vector b. More...
 
scs_float *RESTRICT c
 The (possibly normalized) vector c. More...
 
scs_float *RESTRICT R
 Fixed-point residual R_k. More...
 
scs_float *RESTRICT R_prev
 Fixed-point residual (FPR) of the previous iteration R_{k-1}. More...
 
scs_float *RESTRICT dir
 Direction d_k. More...
 
scs_float *RESTRICT H
 Hessian approximation used by the full Broyden method. More...
 
scs_float *RESTRICT dut
 Direction corresponding to \tilde{u}. More...
 
scs_float *RESTRICT wu
 Variable w_u. More...
 
scs_float *RESTRICT wu_t
 Variable \tilde{w}_u. More...
 
scs_float *RESTRICT wu_b
 Variable \bar{w}_u. More...
 
scs_float *RESTRICT Rwu
 Vector R(w_u) from line search. More...
 
scs_float nrmR_con
 \|Ru_k\|. More...
 
scs_float *RESTRICT Sk
 s_k = u_k - u_{k-1} More...
 
scs_float *RESTRICT Yk
 y_k = R_k - R_{k-1} More...
 
scs_float stepsize
 The current stepsize t_k. More...
 
scs_float *RESTRICT s_b
 Variable that corresponds to the primal slack for the 2nd step of DRS. More...
 
scs_float gTh
 Auxiliary variable. More...
 
scs_float sc_b
 Scaling factor corresponding to b. More...
 
scs_float sc_c
 Scaling factor corresponding to c. More...
 
scs_float nm_b
 Norm of b. More...
 
scs_float nm_c
 Norm of c. More...
 
scs_float kap_b
 Variable for certificates of infeasibility/unboudedness. More...
 
ScsAMatrixA
 The (possibly normalized) A matrix. More...
 
ScsPrivWorkspace *RESTRICT p
 struct populated by linear system solver More...
 
ScsSettings *RESTRICT stgs
 contains solver settings specified by user More...
 
ScsScaling *RESTRICT scal
 contains the re-scaling data More...
 
ScsConeWork *RESTRICT coneWork
 workspace for the cone projection step More...
 
ScsDirectionCache *RESTRICT direction_cache
 A cache for the computation of Broyden or Anderson's acceleration. More...
 

Detailed Description

Workspace for SCS.

Field Documentation

The (possibly normalized) A matrix.

The (possibly normalized) vector b.

The (possibly normalized) vector c.

ScsConeWork* RESTRICT coneWork

workspace for the cone projection step

Direction d_k.

ScsDirectionCache* RESTRICT direction_cache

A cache for the computation of Broyden or Anderson's acceleration.

Dual residual vector.

\text{dr} = A'y + c \tau,

and in SuperSCS, y is \bar{y} and \tau is \bar{\tau}.

Direction corresponding to \tilde{u}.

Auxiliary variable of dimension l-1 used in projections on linear spaces.

scs_float gTh

Auxiliary variable.

Auxiliary variable used in projections on linear spaces.

Hessian approximation used by the full Broyden method.

See Also
full_broyden
scs_float kap_b

Variable for certificates of infeasibility/unboudedness.

Dimension of u_k, that is l = m + n + 1 .

Row dimension of A.

Column dimension of A.

scs_float nm_b

Norm of b.

scs_float nm_c

Norm of c.

scs_float nrmR_con

\|Ru_k\|.

struct populated by linear system solver

Primal residual vector.

\text{pr} = A x + s - b \tau,

and in SuperSCS, x is \bar{x}, s is \bar{s} and \tau is \bar{\tau}

Fixed-point residual R_k.

Fixed-point residual (FPR) of the previous iteration R_{k-1}.

Vector R(w_u) from line search.

Variable that corresponds to the primal slack for the 2nd step of DRS.

scs_float sc_b

Scaling factor corresponding to b.

scs_float sc_c

Scaling factor corresponding to c.

contains the re-scaling data

s_k = u_k - u_{k-1}

scs_float stepsize

The current stepsize t_k.

contains solver settings specified by user

Vector u_k=(x_k,y_k,\tau_k).

Vector \bar{u}_k.

Vector u_{k-1} of the previous iteration.

Vector \tilde{u}_k.

Vector v_k = Qu_k (used only in SCS, not in SuperSCS).

Variable w_u.

Variable \bar{w}_u.

Variable \tilde{w}_u.

y_k = R_k - R_{k-1}


The documentation for this struct was generated from the following file: