# Class: Optimization _Configuration for parameter optimization. Inherits single-stage fields from OptimizationStage. For multi-stage workflows, use 'stages' (ignores inherited single-stage fields). Loss equation references observations directly by name._
URI: [tvbo:Optimization](https://w3id.org/tvbo/Optimization) ```{mermaid} classDiagram class Optimization click Optimization href "./Optimization.html" OptimizationStage <|-- Optimization click OptimizationStage href "./OptimizationStage.html" Optimization : algorithm Optimization --> "0..1" String : algorithm click String href "../http://www.w3.org/2001/XMLSchema#string.html" Optimization : depends_on Optimization --> "0..1" String : depends_on click String href "../http://www.w3.org/2001/XMLSchema#string.html" Optimization : description Optimization --> "0..1" String : description click String href "../http://www.w3.org/2001/XMLSchema#string.html" Optimization : execution Optimization --> "0..1" ExecutionConfig : execution click ExecutionConfig href "./ExecutionConfig.html" Optimization : free_parameters Optimization --> "*" FreeParameter : free_parameters click FreeParameter href "./FreeParameter.html" Optimization : freeze_parameters Optimization --> "*" Parameter : freeze_parameters click Parameter href "./Parameter.html" Optimization : hyperparameters Optimization --> "*" Parameter : hyperparameters click Parameter href "./Parameter.html" Optimization : integration Optimization --> "0..1" Integrator : integration click Integrator href "./Integrator.html" Optimization : label Optimization --> "0..1" String : label click String href "../http://www.w3.org/2001/XMLSchema#string.html" Optimization : learning_rate Optimization --> "0..1" Float : learning_rate click Float href "../http://www.w3.org/2001/XMLSchema#float.html" Optimization : loss Optimization --> "0..1" FunctionCall : loss click FunctionCall href "./FunctionCall.html" Optimization : max_iterations Optimization --> "0..1" Integer : max_iterations click Integer href "../http://www.w3.org/2001/XMLSchema#integer.html" Optimization : name Optimization --> "1" String : name click String href "../http://www.w3.org/2001/XMLSchema#string.html" Optimization : stages Optimization --> "*" OptimizationStage : stages click OptimizationStage href "./OptimizationStage.html" Optimization : warmup_from Optimization --> "0..1" OptimizationStage : warmup_from click OptimizationStage href "./OptimizationStage.html" ``` ## Inheritance * [OptimizationStage](../classes/OptimizationStage.qmd) * **Optimization** ## Class Properties | Property | Value | | --- | --- | | Class URI | [tvbo:Optimization](https://w3id.org/tvbo/Optimization) | ## Slots | Name | Cardinality and Range | Description | Inheritance | | --- | --- | --- | --- | | [execution](../slots/execution.qmd) | 0..1
[ExecutionConfig](../classes/ExecutionConfig.qmd) | Per-optimization execution configuration (overrides experiment-level defaults... | direct | | [integration](../slots/integration.qmd) | 0..1
[Integrator](../classes/Integrator.qmd) | Integration settings for optimization simulations (overrides experiment defau... | direct | | [loss](../slots/loss.qmd) | 0..1
[FunctionCall](../classes/FunctionCall.qmd) | Loss function call | direct | | [stages](../slots/stages.qmd) | *
[OptimizationStage](../classes/OptimizationStage.qmd) | Ordered list of optimization stages | direct | | [depends_on](../slots/depends_on.qmd) | 0..1
[xsd:string](http://www.w3.org/2001/XMLSchema#string) or 
[Algorithm](../classes/Algorithm.qmd) or 
[Exploration](../classes/Exploration.qmd) | Upstream task whose result seeds this optimization | direct | | [name](../slots/name.qmd) | 1
[xsd:string](http://www.w3.org/2001/XMLSchema#string) | Globally unique identifier for the entity | [OptimizationStage](../classes/OptimizationStage.qmd) | | [label](../slots/label.qmd) | 0..1
[xsd:string](http://www.w3.org/2001/XMLSchema#string) | | [OptimizationStage](../classes/OptimizationStage.qmd) | | [description](../slots/description.qmd) | 0..1
[xsd:string](http://www.w3.org/2001/XMLSchema#string) | | [OptimizationStage](../classes/OptimizationStage.qmd) | | [free_parameters](../slots/free_parameters.qmd) | *
[FreeParameter](../classes/FreeParameter.qmd) | Parameters to optimize in this stage | [OptimizationStage](../classes/OptimizationStage.qmd) | | [algorithm](../slots/algorithm.qmd) | 0..1
[xsd:string](http://www.w3.org/2001/XMLSchema#string) | Optimizer for this stage: 'adam', 'adamw', 'sgd', etc | [OptimizationStage](../classes/OptimizationStage.qmd) | | [learning_rate](../slots/learning_rate.qmd) | 0..1
[xsd:float](http://www.w3.org/2001/XMLSchema#float) | | [OptimizationStage](../classes/OptimizationStage.qmd) | | [max_iterations](../slots/max_iterations.qmd) | 0..1
[xsd:integer](http://www.w3.org/2001/XMLSchema#integer) | | [OptimizationStage](../classes/OptimizationStage.qmd) | | [hyperparameters](../slots/hyperparameters.qmd) | *
[Parameter](../classes/Parameter.qmd) | Stage-specific hyperparameters (e | [OptimizationStage](../classes/OptimizationStage.qmd) | | [freeze_parameters](../slots/freeze_parameters.qmd) | *
[Parameter](../classes/Parameter.qmd) | Parameters from previous stages to freeze (keep at optimized value but not up... | [OptimizationStage](../classes/OptimizationStage.qmd) | | [warmup_from](../slots/warmup_from.qmd) | 0..1
[OptimizationStage](../classes/OptimizationStage.qmd) | Previous stage to initialize from | [OptimizationStage](../classes/OptimizationStage.qmd) | ## Usages | used by | used in | type | used | | --- | --- | --- | --- | | [SimulationExperiment](../classes/SimulationExperiment.qmd) | [optimizations](../slots/optimizations.qmd) | range | [Optimization](../classes/Optimization.qmd) | ## Identifier and Mapping Information ### Schema Source * from schema: https://w3id.org/tvbo ## Mappings | Mapping Type | Mapped Value | | --- | --- | | self | tvbo:Optimization | | native | tvbo:Optimization | ## LinkML Source ### Direct
```yaml name: Optimization description: Configuration for parameter optimization. Inherits single-stage fields from OptimizationStage. For multi-stage workflows, use 'stages' (ignores inherited single-stage fields). Loss equation references observations directly by name. from_schema: https://w3id.org/tvbo is_a: OptimizationStage attributes: execution: name: execution description: Per-optimization execution configuration (overrides experiment-level defaults). Useful for setting random_seed, precision, or hardware for optimization phase. from_schema: https://w3id.org/tvbo domain_of: - Analysis - Optimization - Exploration - Inference - Algorithm - Continuation - SimulationExperiment range: ExecutionConfig inlined: true integration: name: integration description: Integration settings for optimization simulations (overrides experiment defaults). If specified, creates a fresh model_fn and state with prepare() before optimization. Can specify different duration, step_size, method than the experiment. If not specified, uses experiment-level integration settings. from_schema: https://w3id.org/tvbo domain_of: - Network - Optimization - Inference - SimulationExperiment range: Integrator inlined: true loss: name: loss description: 'Loss function call. Uses FunctionCall to either: 1. Reference existing function: function: rmse 2. Inline callable: callable: {module: ..., name: ...} Arguments specify inputs (simulated_fc, empirical_fc, etc.)' from_schema: https://w3id.org/tvbo rank: 1000 domain_of: - Optimization range: FunctionCall inlined: true stages: name: stages description: Ordered list of optimization stages. Stages run sequentially. Stage n+1 starts from optimized values of stage n. When defined, inherited single-stage fields are ignored. from_schema: https://w3id.org/tvbo rank: 1000 domain_of: - Optimization - Algorithm range: OptimizationStage multivalued: true inlined: true inlined_as_list: true depends_on: name: depends_on description: 'Upstream task whose result seeds this optimization. An `Algorithm` supplies a single tuned state (start from it). An `Exploration` with an adaptive `strategy` (e.g. nsga2) supplies a *front* of N non-dominated solutions: the optimization then runs batched-parallel — one gradient refinement per Pareto seed — reusing `parallel_mode`/`n_parallel`. If absent, optimization starts from the initial simulation state. Forward-compatible with the unified `Search.depends_on`.' from_schema: https://w3id.org/tvbo rank: 1000 domain_of: - Optimization - Inference - Algorithm inlined: false any_of: - range: Algorithm - range: Exploration class_uri: tvbo:Optimization ```
### Induced
```yaml name: Optimization description: Configuration for parameter optimization. Inherits single-stage fields from OptimizationStage. For multi-stage workflows, use 'stages' (ignores inherited single-stage fields). Loss equation references observations directly by name. from_schema: https://w3id.org/tvbo is_a: OptimizationStage attributes: execution: name: execution description: Per-optimization execution configuration (overrides experiment-level defaults). Useful for setting random_seed, precision, or hardware for optimization phase. from_schema: https://w3id.org/tvbo owner: Optimization domain_of: - Analysis - Optimization - Exploration - Inference - Algorithm - Continuation - SimulationExperiment range: ExecutionConfig inlined: true integration: name: integration description: Integration settings for optimization simulations (overrides experiment defaults). If specified, creates a fresh model_fn and state with prepare() before optimization. Can specify different duration, step_size, method than the experiment. If not specified, uses experiment-level integration settings. from_schema: https://w3id.org/tvbo owner: Optimization domain_of: - Network - Optimization - Inference - SimulationExperiment range: Integrator inlined: true loss: name: loss description: 'Loss function call. Uses FunctionCall to either: 1. Reference existing function: function: rmse 2. Inline callable: callable: {module: ..., name: ...} Arguments specify inputs (simulated_fc, empirical_fc, etc.)' from_schema: https://w3id.org/tvbo rank: 1000 owner: Optimization domain_of: - Optimization range: FunctionCall inlined: true stages: name: stages description: Ordered list of optimization stages. Stages run sequentially. Stage n+1 starts from optimized values of stage n. When defined, inherited single-stage fields are ignored. from_schema: https://w3id.org/tvbo rank: 1000 owner: Optimization domain_of: - Optimization - Algorithm range: OptimizationStage multivalued: true inlined: true inlined_as_list: true depends_on: name: depends_on description: 'Upstream task whose result seeds this optimization. An `Algorithm` supplies a single tuned state (start from it). An `Exploration` with an adaptive `strategy` (e.g. nsga2) supplies a *front* of N non-dominated solutions: the optimization then runs batched-parallel — one gradient refinement per Pareto seed — reusing `parallel_mode`/`n_parallel`. If absent, optimization starts from the initial simulation state. Forward-compatible with the unified `Search.depends_on`.' from_schema: https://w3id.org/tvbo rank: 1000 owner: Optimization domain_of: - Optimization - Inference - Algorithm range: string inlined: false any_of: - range: Algorithm - range: Exploration name: name: name description: Globally unique identifier for the entity. from_schema: https://w3id.org/tvbo/common slot_uri: schema:name identifier: true owner: Optimization domain_of: - BrainAtlas - CommonCoordinateSpace - ParcellationEntity - StudyLayout - StudyTemplate - StudyDirectory - StudyFile - DBSProtocol - ClinicalScale - ClinicalScore - SoftwarePackage - SoftwareRequirement - SoftwareEnvironment - Figure - Event - Tractogram - ResultEntity - Activity - CachedArray - MeasureSpec - NamedArray - GraphGenerator - Binding - ProcedureStep - File - StateValue - Observation - Dynamics - StateVariable - Distribution - Parameter - CouplingInput - Argument - Function - FunctionCall - Callable - DerivedParameter - DerivedVariable - DataSource - OptimizationStage - Exploration - Inference - Prior - Likelihood - UpdateRule - Algorithm - Option - BranchSwitch - Continuation - Coupling - EnvironmentVariable - SchedulerDirective - Reducer range: string required: true label: name: label from_schema: https://w3id.org/tvbo rank: 1000 slot_uri: rdfs:label owner: Optimization domain_of: - ParcellationTerminology - Subject - Session - Dataset - Contact - SoftwareEnvironment - Figure - Panel - Inset - Layer - Colorbar - Rule - Equation - Stimulus - Event - Parcellation - Tractogram - Matrix - Phenotype - Network - Node - Edge - Observation - Dynamics - StateVariable - Parameter - Function - FunctionCall - DerivedVariable - RandomStream - DataSource - OptimizationStage - Exploration - ExplorationAxis - FreeParameter - Inference - AlgorithmStage - TuningObjective - Continuation - Coupling - RegionMapping - SimulationExperiment - SimulationStudy - TimeSeries - NDArray - SpatialDomain - Mesh - SpatialField - FieldStateVariable - DifferentialOperator - BoundaryCondition - PDESolver - PDE range: string description: name: description from_schema: https://w3id.org/tvbo rank: 1000 slot_uri: dcterms:description owner: Optimization domain_of: - CommonCoordinateSpace - Dataset - StudyLayout - StudyTemplate - StudyDirectory - StudyFile - ClinicalScore - SoftwarePackage - SoftwareRequirement - SoftwareEnvironment - Figure - Panel - Equation - Stimulus - Event - Tractogram - Matrix - ResultEntity - Activity - Phenotype - MeasureSpec - NamedArray - Network - GraphGenerator - ProcedureStep - File - Node - Edge - Observation - Dynamics - StateVariable - Parameter - CouplingInput - Argument - Function - FunctionCall - Callable - DerivedParameter - DerivedVariable - RandomStream - DataSource - OptimizationStage - Exploration - ExplorationAxis - FreeParameter - Inference - Prior - Likelihood - UpdateRule - AlgorithmStage - TuningObjective - Algorithm - BranchSwitch - Continuation - Integrator - Coupling - RegionMapping - SimulationExperiment - Study - ResultBinding - TimeSeries - NDArray - SpatialDomain - Mesh - SpatialField - FieldStateVariable - BoundaryCondition - PDESolver - PDE - Reducer range: string free_parameters: name: free_parameters description: Parameters to optimize in this stage. Each entry is a FreeParameter that references an existing Parameter by dotted scope (e.g. "ReducedWongWang.w" or "FastLinearCoupling.G") and supplies optimization-specific metadata (heterogeneous, shape, bounds). No new Parameter is created here. from_schema: https://w3id.org/tvbo rank: 1000 owner: Optimization domain_of: - OptimizationStage - Continuation range: FreeParameter multivalued: true inlined: true inlined_as_list: true algorithm: name: algorithm description: 'Optimizer for this stage: ''adam'', ''adamw'', ''sgd'', etc.' from_schema: https://w3id.org/tvbo ifabsent: string(adam) owner: Optimization domain_of: - Digest - OptimizationStage - AlgorithmInclude - Continuation range: string learning_rate: name: learning_rate from_schema: https://w3id.org/tvbo rank: 1000 ifabsent: float(0.001) owner: Optimization domain_of: - OptimizationStage - Algorithm range: float max_iterations: name: max_iterations from_schema: https://w3id.org/tvbo rank: 1000 ifabsent: integer(100) owner: Optimization domain_of: - OptimizationStage range: integer hyperparameters: name: hyperparameters description: Stage-specific hyperparameters (e.g., b2=0.9999 for adam) from_schema: https://w3id.org/tvbo rank: 1000 owner: Optimization domain_of: - OptimizationStage - Algorithm range: Parameter multivalued: true inlined: true inlined_as_list: true freeze_parameters: name: freeze_parameters description: Parameters from previous stages to freeze (keep at optimized value but not update) from_schema: https://w3id.org/tvbo rank: 1000 owner: Optimization domain_of: - OptimizationStage range: Parameter multivalued: true inlined: false warmup_from: name: warmup_from description: Previous stage to initialize from. Final values from that stage become initial values for this stage. from_schema: https://w3id.org/tvbo rank: 1000 owner: Optimization domain_of: - OptimizationStage range: OptimizationStage inlined: false class_uri: tvbo:Optimization ```