Changed

• Hide the Expert tab in MainWindow UI
• Updated default preset button names in multi-objective sampler settings from product/tool-specific labels to behavior-based labels:
  • Exploitation Default
  • Exploration Default
  • Applied to HypE, NSGA-II, and SPEA-II settings pages
• Updated initial/default selection values in HypE, NSGA-II, SPEA-II, MOEA/D, and NSGA-III settings to make crossover/mutation behavior more explicit and consistent

Fixed

• Optuna-Dashboard Rhino3dm visualization error

v1.4.0 update info

Added

• Check agreement to Terms of Use on startup
  • User must agree before the UI opens
• NEW sampler support
  • Swarm Intelligence
    • PSO (Particle Swarm Optimization)
      • PSO sampler can use community mode
    • GreyWolf (Grey Wolf Optimization)
    • Whale (Whale Optimization Algorithm)
  • Bayesian Optimization
    • TuRBO (Bayesian Optimization with Trust Regions)
    • CARBO (Constrained robust Bayesian optimization of expensive noisy black-box functions with guaranteed regret bounds)
    • Robust GP (Robust Bayesian Optimization under Input Noise)
  • Evolution Algorithm
    • SPEA-II (Strength Pareto Evolutionary Algorithm II)
    • HypE (Hypervolume Estimation Algorithm)
      • These 2 samplers can use community mode
• Sampler variable type handling
  • Verified whether the sampler supports continuous or categorical variables and provided appropriate variable suggestions
• Show Confirmation whether to save optimization results for InMemory mode
  • When using InMemory mode, a confirmation dialog appears to ask whether to save the results after optimization is complete
• Support for local license
  • This is particularly useful for users in environments with restricted internet access or for those who prefer to manage their licenses locally.

Changed

• Disable the second Live Chart when there is only one objective function
  • The second chart's checkbox is unchecked by default when single-objective optimization is used
• Made it clear that the UI's Timeout corresponds to the Study's Timeout
• The display category for samplers like Random has changed from Misc to Exploration in the UI
• HEBO supports multi-objective optimization and Human-in-the-loop optimization
• Significantly improved Python environment performance
  • By implementing a singleton pattern and eliminating the need to repeatedly launch Python environments, performance has been dramatically enhanced
• Including optunahub's samplers
  • Eliminated the need for git or internet connectivity, improving performance.
• Removed ForceReload option

Fixed

• Null exception when handling large numbers of attributes
• Race condition in AttributeManager, ObjectiveManager, and ArtifactManager where values could become stale or empty due to concurrent access from the UI thread and worker thread during optimization
  • Cached manager results so worker thread reads from cache instead of accessing Grasshopper UI objects directly
  • Added IsNewSolutionCompleted flag to GrasshopperManager to ensure worker thread waits for all GetValues to complete after recalculation
  • Fixed SolutionManager.RecalculateAsync to subscribe to SolutionEnd event before triggering NewSolution, preventing a potential deadlock when the solution completes synchronously
• When the rhp license check encounters a load error, an exception occurs that prevents the Tunny UI from launching
  • To address this, the behavior has been modified to skip the license check and instead launch in Community Edition
• To avoid cases where the objective function value cannot be obtained under race conditions, we implemented a retry mechanism if the value is null.