CHANGELOG.md 3.55 KB
Newer Older
Jens Saak's avatar
Jens Saak committed
1 2 3 4 5 6 7 8 9 10 11
## version 2.0.1

### Changed
- many function headers and help texts got improved/completed

### Fixed
- DAE_1 usfs failed for certain systems with non-symmetric A.
- LTV BDF could break in certain situations and was not following the
  general naming scheme for some variables.
- mess_res2_norms would break when more than 4 output arguments were requested

12
## version 2.0
Jens Saak's avatar
Jens Saak committed
13 14

### Added
15 16 17 18 19
- New RADI iteration for AREs
- New splitting methods for autonomous DREs
- New splitting and BDF methods for non-autonomous DREs
- New operator manager only requires non-empty functions and replaces
  non-existent ones with a general `mess_do_nothing` function 
Jens Saak's avatar
Jens Saak committed
20 21 22 23 24
- CI testing
  - demos serve as system tests 
  - additional unit tests for the smaller building blocks and backend routines

### Changed
25
- improved Riccati iteration
Jens Saak's avatar
Jens Saak committed
26
- updated minimum required/recommended Matlab and Octave versions
27 28 29
  (see `DEPENDENCIES.md`)
- unified function interfaces for top level calls
- unified handling of low rank updated operators. Now always A+UV' is
Jens Saak's avatar
Jens Saak committed
30
  used. (Note the sign of the update and the transposition in V)
31
- major updates in the MOR routines
Jens Saak's avatar
Jens Saak committed
32
- some restructuring in the opts structure.
33 34 35 36 37
  * `opts.adi.shifts` has moved to `opts.shifts` such that also RADI
    can use it independent of ADI
  *  opts.norm now determines the norm for all methods rather than
     having to consistently specifiy the same norm in each substructure
  * initial feedbacks for the Riccati solvers are now stored in the
Jens Saak's avatar
Jens Saak committed
38 39
    `opts` structure for the method rather than `eqn`
- The projection shift routine uses the flag `opts.shifts.implicitVtAV`.
40
  Default is `true`. If set to `false` A*V is computed explicitly.
41 42 43
- redesign of the demos
  - turned scripts into actual demo functions
  - new demos for indefinite AREs and H-infinity control
Jens Saak's avatar
Jens Saak committed
44 45 46 47 48 49 50 51 52 53 54 55 56 57

### Fixed
- several consistency updates and bug fixes
- general code cleaning and pretty printing

## version 1.0.1

### Changed
- updated documentation 
- Removed replacements directory since its content was not needed for
  Matlab after release 2010b and Octave after 4.0. 

### Fixed
- Minor consistency and bug fixes and improved integrity of metafiles.
58
- CI testing
Jens Saak's avatar
Jens Saak committed
59
  - demos serve as system tests
60 61
  - additional unit tests for the smaller building blocks and backend routines

Jens Saak's avatar
Jens Saak committed
62 63
## version 1.0.1
- Minor consistency and bug fixes and improved integrity of metafiles.
Jens Saak's avatar
Jens Saak committed
64
- updated documentation
65
- Removed replacements directory since its content was not needed for
Jens Saak's avatar
Jens Saak committed
66 67
  Matlab after release 2010b and Octave after 4.0.

Jens Saak's avatar
Jens Saak committed
68

Jens Saak's avatar
Jens Saak committed
69 70
## version 1.0
Compared to the predecessor lyapack a couple of things have changed.
Jens Saak's avatar
Jens Saak committed
71

Jens Saak's avatar
Jens Saak committed
72 73 74
- The user supplied functions are now managed by an operator manager
- The low rank ADI now has:
  - optimized treatment of E matrices in generalized equations
75
  - more choices for shift selection, including completely automatic
Jens Saak's avatar
Jens Saak committed
76
    generation of shifts
Jens Saak's avatar
Jens Saak committed
77 78 79 80 81 82 83 84 85
  - improved stopping criteria based on low rank factors of the current residual
  - automatic generation of real low rank factors also for complex shifts
- The Newton-Kleinman iteration features:
  - optimized treatment of E matrices in generalized equations
  - improved stopping criteria based on low rank factors of the current residual
  - inexact Newton, line search and Galerkin projection acceleration
- Examples have been extended
- The Riccati iteration for H-infinity Riccati equations was added
- DSPMR has not yet been ported to the new infrastructure
86 87 88
- The SRM routine for balanced truncation is only available for
  none-DAE systems. Still, DAE versions are included in the
  corresponding DEMOS. 
Jens Saak's avatar
Jens Saak committed
89 90
- A tangential IRKA implementation for non-DAE systems was added