v0.5.3

Hypatia v0.5.3

Diff since v0.5.2

Closed issues:

• run some JuMP example instances with non-Float64 reals (#598)

v0.5.2

Hypatia v0.5.2

Diff since v0.5.1

Merged pull requests:

• fix ambiguity on nightly (#745) (@chriscoey)
• prepare experiments using spectral function cones (#746) (@lkapelevich)
• add NegPower and NegSqrt separable spectral functions, remove Inv (#747) (@chriscoey)
• add ECOS to natvsext benchmarks, improve scripts (#748) (@lkapelevich)
• fix analyze formulation aggregation (#749) (@chriscoey)
• add explicit inverse Hessian for powermean (#750) (@lkapelevich)
• remove warn_attributes option (#751) (@chriscoey)
• WIP: inverse Hessian for generalized power cone (#752) (@lkapelevich)
• simplify inv hess oracles for sepspectral cone (#755) (@lkapelevich)
• cleanup log and logdet cones (#756) (@lkapelevich)
• cleanup rootdet/geomean (#757) (@chriscoey)
• update benchmarks; adjust geomean/rtdet/log/logdet constants (#758) (@chriscoey)
• add some complex model utils (#759) (@chriscoey)
• Update natvsext benchmarks README.md (#760) (@lkapelevich)
• adjust benchmark list (#761) (@chriscoey)
• global var to constant (#762) (@matbesancon)
• DNMY: update benchmark script for wsos examples (#763) (@lkapelevich)
• rename neggeom to negrtdet; adjust bench sizes (#764) (@chriscoey)
• decrease memory overhead in model construction for expdesign (#765) (@chriscoey)
• make objective more reasonable in size in polynorm example (#766) (@lkapelevich)
• update readmes for benchmarks (#767) (@lkapelevich)
• improve speed of separable spectral cone oracles (#768) (@chriscoey)
• update stepper benchmarking instructions (#770) (@lkapelevich)
• improve spectral cone oracles, especially inv hess (#771) (@chriscoey)

v0.5.1

Hypatia v0.5.1

Diff since v0.5.0

Merged pull requests:

• fix proximity aggregation (#743) (@chriscoey)
• fix bug in predorcent phase selection (#744) (@chriscoey)

v0.5.0

Hypatia v0.5.0

Diff since v0.4.0

Closed issues:

• try using higher order Runge–Kutta methods for prediction (#51)
• replace errors with appropriate statuses (#329)
• check that SuiteSparse caches are using the maximum number of threads (#361)

Merged pull requests:

• WIP: add more cone examples (#686) (@lkapelevich)
• add two more QRE cone examples (#687) (@chriscoey)
• add missing examples to stepper benchmarks (#689) (@lkapelevich)
• add some separable spectral cone examples (#690) (@chriscoey)
• generalize experiment design examples (#691) (@chriscoey)
• make minor changes to run benchmarks (#692) (@lkapelevich)
• run nightly and fix strange stack overflow error, only use cholmod sparsepsd on recent julia, update portfolio example (#693) (@chriscoey)
• fix extender names (#694) (@chriscoey)
• don't add redundant constraints in example (#695) (@chriscoey)
• clean up some style and infrastructure (#696) (@chriscoey)
• WIP: clean up cones and linear algebra (#697) (@chriscoey)
• reduce allocations for (inv) hessians, more clean up (#698) (@chriscoey)
• clean up polynomial interpolations and polymin (#699) (@chriscoey)
• WIP: style fixes and coverage improvements (#700) (@chriscoey)
• add comments and fix bug for TOA without curve (#701) (@lkapelevich)
• CompatHelper: add new compat entry for "SpecialFunctions" at version "1.4" (#702) (@github-actions[bot])
• rename correction to adjustment/dder3 for clarity (#703) (@chriscoey)
• clean up example instances (#704) (@chriscoey)
• WIP: do some matrix factorizations/inverses in-place (#705) (@chriscoey)
• add various instances to sparseLMI and fix rosenbrock data (#706) (@lkapelevich)
• WIP: add docs using Documenter (#707) (@chriscoey)
• WIP: rename and refactor math in log cones (#708) (@lkapelevich)
• add to docs and edit readme (#709) (@chriscoey)
• CompatHelper: add new compat entry for "DocStringExtensions" at version "0.8" (#710) (@github-actions[bot])
• more docs (#711) (@chriscoey)
• update docs structure and readme (#712) (@chriscoey)
• improve docs and add docstrings (#713) (@chriscoey)
• WIP: rename and refactor geomean-like cones (#714) (@lkapelevich)
• improve scripts, put examples in a module, check for possible duplicate instances (#715) (@chriscoey)
• WIP: speed up some oracles (#716) (@chriscoey)
• add more comments in analysis script (#717) (@lkapelevich)
• WIP: improve matrix rel entr cone oracles (#718) (@chriscoey)
• add example with powermean cone (#719) (@lkapelevich)
• clean up geomean, use ordereddict to run examples (#720) (@chriscoey)
• make minor edits to example descriptions (#721) (@lkapelevich)
• simplify directions for kappa part (#722) (@chriscoey)
• WIP: refactor terms in some oracles (#723) (@chriscoey)
• describe cones interface and scripts more in docs (#724) (@chriscoey)
• fix data not being added into ordereddict (#725) (@lkapelevich)
• use ordered dicts for example instances (#726) (@chriscoey)
• WIP: allow a more SY-like predorcent stepper (#727) (@chriscoey)
• minor cleanup of generalized power cone (#728) (@chriscoey)
• cleanup dense linear algebra and fix nondeterministic traces due to hessian factorization switches (#729) (@chriscoey)
• add automated extended formulations for separable spectral cone examples (#730) (@chriscoey)
• add more spectral cone example formulations, refactor extended formulations (#731) (@chriscoey)
• add example tests and some dual formulations for spectral cones (#732) (@chriscoey)
• update steppers to run in benchmarks (#733) (@chriscoey)
• make sep spec tests more robust and add some instances (#734) (@chriscoey)
• fix qr deprecation warning (#735) (@chriscoey)
• update scripts to run all extenders (#736) (@chriscoey)
• clean up qrchol LHS setup and use a better heuristic for sqrt hess oracles (#737) (@chriscoey)
• update instances hitting time limits in benchmarks (#738) (@lkapelevich)
• solve larger doptimaldesign benchmark instances (#739) (@chriscoey)
• minor natvsext script fixes (#740) (@chriscoey)
• loosen tols if possible for instances where no stepper converges and updates to analysis (#741) (@lkapelevich)
• change the way caps are calculate in stepper analysis script (#742) (@lkapelevich)

v0.4.0

Hypatia v0.4.0

Diff since v0.3.0

Closed issues:

• return SlowProgress if some residual norms start increasing too much (#327)
• tighten tols for infeasibility detection check (#347)
• improve termination criteria for outer 6x6 iterative refinement (#375)
• fix argument types of vec to mat functions (#392)
• optimize LMI cone barrier evals (#477)
• for WSOS cones, special case P_r for which L_r = 1 (just use log of 1x1 Lambda instead of logdet) (#533)
• try checking hessian nbhd after heuristic nbhd is satisfied for all cones (#534)
• experiment with higher order correctors (again) (#567)
• MOI interface: set_silent not respected (#640)
• get a more efficient TOA for lmi cone (#654)

Merged pull requests:

• add matrix relative entropy cone (#574) (@lkapelevich)
• fix type issues with interpolation functions (#629) (@chriscoey)
• rename predorcent stepper file (#630) (@chriscoey)
• improve predorcent stepper (#631) (@chriscoey)
• improve predorcent stepper and use curve search (#633) (@chriscoey)
• add more entropy-like cones (#634) (@lkapelevich)
• CompatHelper: bump compat for "IterativeSolvers" to "0.9" (#635) (@github-actions[bot])
• cleanup and add barrier oracles numerics checks (#637) (@chriscoey)
• cleanup and improve slow progress checks (#638) (@chriscoey)
• add third order derivatives for relative entropy (#639) (@lkapelevich)
• add and test stepper options and cleanup search (#641) (@chriscoey)
• if not spawning in runbenchmarks, run compile instance (#642) (@chriscoey)
• simplify steppers and alpha search, add optional initial alpha heuristic (#643) (@chriscoey)
• delete a redundant line of code (#646) (@lkapelevich)
• improve predorcent stepper (#647) (@chriscoey)
• correct MOI.Silent (#648) (@lkapelevich)
• add timings for subprocedures (#649) (@chriscoey)
• use lower bound on hessian nbhd to speed up search (#650) (@chriscoey)
• time factorizations and system solver setup (#652) (@chriscoey)
• remove constant multiplier from rootdet barrier (#653) (@lkapelevich)
• rename randompolymat to nearestpolymat (#655) (@lkapelevich)
• refactor wsos matrix oracles for faster grad (#656) (@lkapelevich)
• refactor wsos L1 cone, making grad faster (#657) (@lkapelevich)
• refactor wsos L2 cone, making grad faster (#658) (@lkapelevich)
• add scripts for benchmarking steppers (#659) (@lkapelevich)
• add various example instances to benchmark (#661) (@lkapelevich)
• get a better LMI TOA (#662) (@lkapelevich)
• faster TOA for wsos nonnegative (#663) (@lkapelevich)
• cleanup cone oracle tests, check oracle allocations (#664) (@chriscoey)
• add power benchmark instances and tweaks in stepper run script (#665) (@lkapelevich)
• speed up matrix wsos TOA (#666) (@lkapelevich)
• WIP: fix some slow oracles and failing instances (#667) (@chriscoey)
• make stepper analysis script produce more CSVs (#668) (@lkapelevich)
• make more tweaks to benchmark analysis scripts (#669) (@lkapelevich)
• improve scalar wsos oracles and test complex case (#670) (@lkapelevich)
• improve LMI cone oracles (#671) (@chriscoey)
• improve wsosmat corrector slightly (#672) (@chriscoey)
• reduce time complexity in the TOO of WSOS L2 cone (#673) (@lkapelevich)
• WIP: use expensive hess prods, only use iterative refinement when needed (#674) (@chriscoey)
• record example type in dataframe (#675) (@chriscoey)
• improve wsos oracles speed (#676) (@chriscoey)
• improve time complexity for WSOS L1 TOO (#677) (@lkapelevich)
• use expensive hessian products only when needed (#678) (@chriscoey)
• count number of instances with loosened tolerances (#679) (@lkapelevich)
• rename shift to back (#680) (@lkapelevich)
• remove unneeded preallocated space in matrix wsos (#681) (@lkapelevich)
• refactor hess prod and correction for wsosmat (#682) (@chriscoey)
• WIP: add different implementation types for sparse PSD cone, add new example (#683) (@chriscoey)
• WIP: add new generic matrix monotone epigraph cone (#684) (@chriscoey)
• delete unused import (#685) (@lkapelevich)

v0.3.0

Hypatia v0.3.0

Diff since v0.2.0

Merged pull requests:

• add postprocessing scripts, various cone and example improvements (#609) (@lkapelevich)
• WIP: add lower complexity oracles for more complicated WSOS cones (#615) (@lkapelevich)
• refactor and simplify arrow hessian oracles (#616) (@chriscoey)
• add benchmark problems for more complicated WSOS cones (#617) (@lkapelevich)
• remove some allocations in wsosl2 corrector (#618) (@lkapelevich)
• add corrector for WSOS L1 cone (#619) (@lkapelevich)
• improve benchmark processing (#620) (@chriscoey)
• set appropriate defaults for sparse vs dense MOI models (#621) (@chriscoey)
• CI: try julia release and nightly (#622) (@chriscoey)
• allow getting nu in postprocessing (#623) (@chriscoey)
• update processbenchmarks for DataFrames changes (#624) (@lkapelevich)
• CompatHelper: bump compat for "LinearMaps" to "3.0" (#625) (@github-actions[bot])
• cleanup examples and process script (#627) (@chriscoey)

v0.2.0

Hypatia v0.2.0

Diff since v0.1.1

Merged pull requests:

• add inbounds annotations in sparse psd (#604) (@lkapelevich)
• make benchmark script work on julia 1.5 (#605) (@chriscoey)
• improve benchmark sets and script (#606) (@chriscoey)
• minor bug fixes and cleanup (#607) (@chriscoey)
• change JuMP portfolio example to rebalancing (#608) (@chriscoey)
• update definitions of violations in processing of examples (#610) (@lkapelevich)
• fix some benchmark script failures (#611) (@chriscoey)
• use new convergence conditions (#612) (@chriscoey)
• improve benchmark script; use more views etc to improve efficiency (#614) (@chriscoey)

v0.1.1

Hypatia v0.1.1

Diff since v0.1.0

Merged pull requests:

• relax version bounds for julia and run julia nightly in travis (#594) (@chriscoey)
• initialization fixes for wsosl1 cone (#595) (@lkapelevich)
• minor tweaks, add missing inbounds (#596) (@lkapelevich)
• test more real types thru MOI tests (#597) (@chriscoey)
• remove DynamicPolynomials dependency and lagrange polys from ModelUtilities (#599) (@chriscoey)
• fix access for some pointers (#600) (@lkapelevich)
• cleanup stepper and fix corrector bug (#601) (@chriscoey)
• add wsosl1 example and fix bugs in wsos l1 cone (#602) (@lkapelevich)
• more stepper cleanup and options (#603) (@chriscoey)

v0.1.0

Initial public release of Hypatia solver for generic conic optimization. Code is undocumented. Expect bugs and major interface changes.

May 2020 experiments

change back sparsity factor

Former-commit-id: e1995146414569f38e3808986af43c9cc50b9a8a