ead95d20c3
perm_avg → perm_avg!
2019-10-31 00:16:06 +01:00
70bf9ad74c
perm → Generic.Perm
2019-10-30 23:58:10 +01:00
dc14b2fd95
replace Settings.warmstart by force_compute
2019-10-30 23:56:53 +01:00
a5a3d9d069
modernize naive sdp formulation
2019-08-02 11:52:19 +02:00
5e821286e5
define and use one(::Group)
2019-07-05 19:23:37 +02:00
4bce106c2d
better logging
2019-07-05 19:23:36 +02:00
e23b6a85dc
include sqadjop.jl from 1812.03456.jl file
2019-07-05 19:23:35 +02:00
cc9d3db846
we act on MatAlgElem now; use MatAlgebra
2019-07-04 22:52:21 +02:00
88b460a959
use NCRing[Elem]s instead of MatSpace and use NCRing in Settings
2019-07-04 22:52:02 +02:00
6c906b05cb
add sqadjop.jl and unit tests
2019-06-30 13:19:24 +02:00
b9dc701f17
when acting on MatElem always act on the lhs columns/rows
2019-06-30 11:58:32 +02:00
4c4ef195e1
replace radius by halfradius where appropriate
2019-06-30 11:56:57 +02:00
8467b0d432
slightly faster perm action on MatElem
2019-06-28 09:15:01 +02:00
761879f5fd
fix rebase
2019-06-28 01:34:37 +02:00
1345257732
further speedups for WreathProduct on Mats
2019-06-28 01:24:04 +02:00
70c72d28d9
further speedups for perm action on Automorphisms
2019-06-28 01:24:03 +02:00
7e854f902a
simplify WreathProductElem action on Automorphisms
2019-06-28 01:24:02 +02:00
55b7ed09bc
reorganize actions
2019-06-28 01:24:01 +02:00
f4f9dfe21d
much faster WreathProductElem actions
2019-06-28 01:24:00 +02:00
4cef591dfa
faster permutation actions on GroupElements
2019-06-28 01:23:53 +02:00
7a8b21db3c
relax type assumptions in compute_SOS and else
2019-06-28 01:20:48 +02:00
b5f12501f9
try/catch file loading
2019-04-17 12:37:41 +02:00
2a84d04edc
test the correctness of hpc sos computation
2019-04-16 17:05:34 +02:00
70c160411d
cosmetic changes (mostly empty lines)
2019-04-12 23:18:48 +02:00
ed910ccde5
rename computeλandP → approximate_by_SOS
2019-04-02 23:04:46 +02:00
bb42538ee5
rename distance_to_positive_cone → certify_SOS_decomposition
...
and make it a bit more general
2019-04-02 23:04:45 +02:00
6fde981b32
save warmstart in time-stamped file;
...
don't override warmstart.jld if there are any nans in the solution (e.g.
solver was interrupted)
2019-04-02 23:04:44 +02:00
60bdbd0e74
optimized version of compue_SOS
...
using accumulator matrix to allow auto-vectorisation
2019-03-15 08:13:38 +01:00
2602d0a181
Revert "Thread.@threads compute_SOS (better precision!)"
...
This reverts commit ad1b11f10d
.
2019-03-14 14:25:32 +01:00
ad1b11f10d
Thread.@threads compute_SOS (better precision!)
2019-03-12 15:47:31 +01:00
7d0677333c
small stylistic tweaks for @constraint(s)
2019-03-10 16:52:54 +01:00
0b92aebb2f
use @info, @warn without ()
2019-02-24 00:17:20 +01:00
d69e33629b
better info in sparsify!
2019-02-21 21:10:36 +01:00
1f5e7ebdbe
fix performance regression in generating constraints
2019-02-21 14:53:11 +01:00
fb8ec1578a
use multi-line @info where possible
2019-02-21 14:52:20 +01:00
a2f8e3dc17
update the language to MOI
...
* use JuMP.OptimizerFactory
* use value(m[:λ]) to retrive values
2019-02-21 14:51:15 +01:00
49c8ffa900
warmstarting is limited to SCS specific hacks
...
until https://github.com/JuliaOpt/SCS.jl/pull/138 is merged
2019-02-21 14:47:39 +01:00
689bae035e
simplify PropertyT.solve due to changes in MOI
2019-02-21 14:45:58 +01:00
5199189e5f
update SOS_problem to JuMP-0.19 syntax
...
thanks to MOI we don't have to pass λ around as it's accesible via
m[:λ];
Ps need to be kept reference of, since it's an anonymous variable
2019-02-21 14:43:38 +01:00
1ce892dc14
workaround for https://github.com/JuliaLang/julia/issues/30739
2019-01-28 18:40:39 +01:00
5ef5b93dbd
fix info
2019-01-28 17:03:24 +01:00
8e4dd2072c
endof -> lastindex
2019-01-28 08:47:40 +01:00
581097b7d3
add: tests for oSL(4,Z); global @info
2019-01-28 08:42:40 +01:00
4768bb6e35
Merge branch 'AutFn' into enh/julia-v0.7
2019-01-17 08:39:13 +01:00
48c43b4c59
minimize imports
2019-01-14 17:46:36 +01:00
80ac135f59
logging changes
2019-01-14 17:46:13 +01:00
af009547b5
fix: distance_to_positive_cone expects HALF the support radius
2019-01-14 17:45:23 +01:00
3e6a1dde51
rework Settings
2019-01-14 17:44:13 +01:00
5492998e7e
group actions definitions to orbitdata.jl
2019-01-11 07:54:36 +01:00
76b8fed6b1
update for changes in AbstractAlgebra, Groups, GroupRings
...
* MatSpaces are no longer Rings, they are modules (that's a hack, I
know)
* DirectProduct → DirectPower
* full → GroupRings.dense
* constructors of GroupRings require explicit basis
2019-01-11 06:37:24 +01:00