Skip to main content

Runtime TargetSpec

class TargetSpec(BaseModel)  # adk.models.optimization.targets
note

Not the same as the model TargetSpec in target_specifications.json. Runtime specs appear in OptimizationContext.targets after the ADK parses the platform payload. You read these during optimization; you do not author them on disk.

Base type for target entries in the optimization context. Subclasses add value and observation bounds for scalar, vector, and matrix shapes.

Definition​

class TargetSpec(BaseModel):
    name: str = Field(min_length=1)
    fn: Fn

Members​

name​

name: str = Field(min_length=1)

The unique identifier or name of the target. Must contain at least one character.

fn​

fn: Fn

The objective function associated with the target. Defines how the optimization process measures and evaluates success or deviation from desired outcomes.

ValueBounds

class ValueBounds(BaseModel)

Defines the lower and upper numeric limits for a single scalar value used in a target specification.

Definition​

class ValueBounds(BaseModel):
    min: float
    max: float

Members​

min​

min: float

Minimum allowable value for the target or observed data.

max​

max: float

Maximum allowable value for the target or observed data.

ScalarTargetSpec

class ScalarTargetSpec(TargetSpec)

Extends TargetSpec to represent a scalar target specification, including bounds for both target and observation values.

Definition​

class ScalarTargetSpec(TargetSpec):
    target_values: ValueBounds
    observation_bounds: ValueBounds

Members​

target_values​

target_values: ValueBounds

The desired target value range that the optimization aims to achieve.

observation_bounds​

observation_bounds: ValueBounds

The permissible range of observed values during optimization for validation and constraint checking.

VectorTargetSpec

class VectorTargetSpec(TargetSpec)

Extends TargetSpec to define vector-based targets, supporting multiple elements each with their own bounds.

Definition​

class VectorTargetSpec(TargetSpec):
    target_values: list[ValueBounds]
    observation_bounds: list[ValueBounds]

Members​

target_values​

target_values: list[ValueBounds]

List of value bounds representing the desired range for each element in the target vector.

observation_bounds​

observation_bounds: list[ValueBounds]

List of value bounds defining acceptable observed ranges for each vector element during optimization.

MatrixTargetSpec

class MatrixTargetSpec(TargetSpec)

Extends TargetSpec to define matrix-based targets for multidimensional optimization objectives.

Definition​

class MatrixTargetSpec(TargetSpec):
    target_values: list[list[ValueBounds]]
    observation_bounds: list[list[ValueBounds]]

Members​

target_values​

target_values: list[list[ValueBounds]]

Nested list of ValueBounds defining the target value range for each cell in the matrix target.

observation_bounds​

observation_bounds: list[list[ValueBounds]]

Nested list of ValueBounds defining the acceptable observed range for each matrix element during optimization.