v2.3.0

GalacticOptim v2.3.0

Diff since v2.2.1

Merged pull requests:

• CI for LTS (#197) (@ChrisRackauckas)
• add using Zygote to tutorial (#202) (@ArnoStrouwen)
• CompatHelper: bump compat for ArrayInterface to 4, (keep existing compat) (#203) (@github-actions[bot])

v2.2.1

GalacticOptim v2.2.1

Diff since v2.2.0

Closed issues:

• Allow callback to stop the optimization (#188)
• UndefVarError: AutoForwardDiff not defined (#193)
• Constraints being ignored (#195)
• LBFGS Stopping criteria (#196)
• solve callback parameter ignored by NLopt solvers (#199)

Merged pull requests:

• Change MultiStartOptimization interface to generic API (#182) (@jonasmac16)
• Update nlopt.md (#186) (@Vaibhavdixit02)
• fix update of theta so minimizer is returned when maxiters is hit (#187) (@m-bossart)
• Clarify return value from a callback in docs (#190) (@Vaibhavdixit02)
• Make use of Flux.Optimise.AbstractOptimiser (#192) (@lungd)
• Better error message for constrained optimizers (#198) (@ChrisRackauckas)
• Fix nlopt's compatibility with callbacks (#200) (@ChrisRackauckas)

v2.2.0

GalacticOptim v2.2.0

Diff since v2.1.0

Closed issues:

• Automatically use Fminbox wrapping if OptimizationProblem has bounds with Optim optimisers (#172)
• Returning minimizing parameters for Flux optimizers when maxiters is reached (#178)
• Handling of box constraints for Optim.ParticleSwarm (#179)

Merged pull requests:

• Use the minimum with save_best in flux optimiser (#180) (@Vaibhavdixit02)
• Fixing and amending the documentation (#181) (@jonasmac16)
• Replaced incompatible emojis in docs (#183) (@jonasmac16)
• Automatically wrap Optim solver in Fminbox (#184) (@jonasmac16)

v2.1.0

GalacticOptim v2.1.0

Diff since v2.0.3

Closed issues:

• Hook sciml_train into MOI (#3)
• Can't get NLopt to work (#70)
• Support MAX_SENSE in the MOI wrapper (#148)
• Max-sense for NLopt and Optim (#155)
• UndefVarError: AutoZygote not defined (#158)
• Cannot set lower and upper bound for SimulatedAnnealing OptimizationProblem (#171)

Merged pull requests:

• Add MOI interface with MOIOptimizationProblem (#142) (@blegat)
• Make MOIOptimizationProblem fields concrete-typed (#144) (@blegat)
• Add sense kwarg to MOI __solve dispatch (#150) (@Vaibhavdixit02)
• Make MOI interface conforming and add a ton of documentation (#153) (@ChrisRackauckas)
• Made NLopt.jl interface properly handle options & minor changes to MOI (#160) (@jonasmac16)
• Harmonised interfaces and added common arguments (#163) (@jonasmac16)
• Fixed MOI warning (#164) (@jonasmac16)
• First implementation of Metaheuristic.jl interface (#165) (@jonasmac16)
• First implementation of Nonconvex.jl (#166) (@jonasmac16)
• Fix typo in mathoptinterface (#168) (@nandoconde)
• Add NLopt maxsense support (#169) (@Vaibhavdixit02)
• Add Optim maxsense and organise rosenbrock.jl (#170) (@Vaibhavdixit02)
• Fixes requires issue (#173) (@jonasmac16)
• New interface to GCMAES.jl (#174) (@jonasmac16)
• New interface to SpeedMapping.jl (#175) (@jonasmac16)
• New basic interface to NOMAD.jl (#176) (@jonasmac16)
• Remove "higher level docs" and put a higher level table (#177) (@ChrisRackauckas)

v2.0.3

GalacticOptim v2.0.3

Diff since v2.0.2

v2.0.2

GalacticOptim v2.0.2

Diff since v2.0.1

v2.0.1

GalacticOptim v2.0.1

Diff since v2.0.0

Closed issues:

• Error requiring Flux from GalacticOptim (#141)

v2.0.0

GalacticOptim v2.0.0

Diff since v1.3.0

Merged pull requests:

• Use maxfuncevals in BBO solve dispatch instead of maxsteps (#138) (@Vaibhavdixit02)
• Generalize Flux to allow use of other AD systems (#139) (@ChrisRackauckas)
• Prune imports (#140) (@ChrisRackauckas)

v1.3.0

GalacticOptim v1.3.0

Diff since v1.2.0

Merged pull requests:

• CompatHelper: bump compat for "Documenter" to "0.27" for package docs (#137) (@github-actions[bot])

v1.2.0

GalacticOptim v1.2.0

Diff since v1.1.0

Closed issues:

• Re-enable BlackBoxOptim test (#111)
• Pre-compile issues (#125)
• minibatch tutorial not working with Fminbox (#133)

Merged pull requests:

• Removed/commented out BBO multiobjective code chunks (#127) (@jonasmac16)
• Update DataLoader syntax (#128) (@Vaibhavdixit02)
• Enable BBO test and remove EnsembleOptimize commented test (#129) (@Vaibhavdixit02)
• Interface simpl (#131) (@lungd)
• Fix train_loader inside minibatching tutorial. 2 args -> tuple (#132) (@lungd)
• Pass args to gradient function in Fminbox/SAMIN dispatch (#134) (@Vaibhavdixit02)