v3.0.1

• Add transfer function module (please read this paper )
• Add two new datatypes: TransferBinaryVar and TransferBoolVar.
• Fix bug un-order variables in PermutationVar data type.
• Update data type of encoded solution in BoolVar data type.
• Update correct function in BoolVar and BinaryVar.
• Fix bug reproduce results in GA, WCA, and EHO optimizers.
• Fix bug higher probability of 0 value in IntegerVar data type.

v3.0.0

Based on our new proposed classes, solving continuous and discrete problems is never that easy.

Add

• space module with: FloatVar, IntegerVar, StringVar, BoolVar, PermutationVar, BinaryVar, and MixedSetVar classes
  • FloatVar: handle problem with solution's format as float value
  • IntegerVar: handle problem with solution's format as integer value
  • StringVar: handle problem with solution's format as string value
  • BoolVar: handle problem with solution's format as boolean value (True or False)
  • PermutationVar: handle problem with solution's format as permutation value
  • BinaryVar: handle problem with solution's format as binary value (0 or 1)
  • MixedSetVar: handle problem with solution's format as mixed discrete set
• target module with Target class contains:
  • objectives (list), weights (list) to calculate fitness, and fitness (number)
• agent module with: Agentte class that is a placeholder for a search agent, it contains at least two attributes:
  • solution (position - np.ndarray), and a target object

Update

• Convert all optimizers to use new classes
• Convert Tuner and MultiTask classes
• Rename all un-official (developed by our team) optimizers to DevOptimizerName
• Update tests and documents
• Update some examples, not all examples have converted yet (utils and applications folders)

v2.5.4

• Remove deepcopy() to improve the computational speed
• Update the parameter's order in Tuner class
• Update the saving's bug when using Termination in Multitask
• Remove ILA optimizer
• Rename "amend_position()" definition in some algorithms to "bounded_position()".
• Add a "amend_position()" function in Optimizer class. This function will call two functions.
  • bounded_position() from optimizer. This means for optimizer level (get in valid range of position)
  • amend_position() from problem. This means for problem level (transform to the correct solution)
• Fix bugs coefficients in GWO-based optimizers.
• Fig bug self.epoch in SCSO optimizer.
• Fix bug self.dyn_pop_size when pop_size is small value
• Move SHADE-based optimizers from DE to SHADE module in evolutionary_based group
• Add Improved Grey Wolf Optimization (IGWO) in GWO algorithm
• Add Tabu Search (TS) to math-based group
• Add get_all_optimizers() and get_optimizer_by_name() in Mealpy
• Rename the OriginalSA to SwarmSA in SA optimizer
• Add the OriginalSA and GaussianSA in SA optimizer
• Update parameters in OriginalHC and SwarmHC
• Update ParameterGrid class to produce the dict with same order as original input
• Add export_figures() to Tuner class. It can draw the hyperparameter tuning process.
• Fix several bugs in docs folders.

v2.5.4-alpha.6

• Fix bug self.dyn_pop_size when pop_size is small value
• Move SHADE-based optimizers from DE to SHADE module in evolutionary_based group
• Add Improved Grey Wolf Optimization (IGWO) in GWO algorithm
• Add Tabu Search (TS) to math-based group
• Add get_all_optimizers() and get_optimizer_by_name() in Mealpy

v2.5.4-alpha.5

• Fig bug self.epoch in SCSO optimizer.

v2.5.4-alpha.4

• Update documents for MultiTask
• Fix bugs coefficients in GWO-based optimizers.

v2.5.4-alpha.3

Release new alpha version

v2.5.4-alpha.2

• Update the parameter's order in Tuner class
• Update the saving's bug when using Termination in Multitask
• Remove ILA optimizer
• Rename "amend_position()" definition in some algorithms to "bounded_position()".
• Add a "amend_position()" function in Optimizer class. This function will call two functions.
  • bounded_position() from optimizer. This means for optimizer level (get in valid range of position)
  • amend_position() from problem. This means for problem level (transform to the correct solution)

v2.5.4-alpha.1

Update

• Update the parameter's order in Tuner class
• Update the saving's bug when using Termination in Multitask

v2.5.3

Update

• Fix bug in roulette-wheel-selection in Optimizer
• Update multitask with input modes and terminations
• Update Tuner with more input parameters
• Add Lévy flight, and the selective opposition version of the artificial rabbit algorithm (LARO)
• Add Modified Gorilla Troops Optimization (MGTO)
• Update Giant Trevally Optimizer as requested by the authors
  • Matlab101GTO: This version was used to produce the results presented in the paper.
  • Matlab102GTO: This is a new version provided by the authors (Matlab link), which has been updated recently to
    reduce computation time.
  • OriginalGTO: This version is implemented exactly as described in the paper.