UQPyL Integration ​

HydroPilot provides UQPyLAdapter to bridge HydroPilot configurations into UQPyL optimization workflows.

Core idea ​

UQPyLAdapter wraps a HydroPilot configuration as a UQPyL Problem. This lets UQPyL optimizers call HydroPilot model evaluations through the standard Problem interface.

What the adapter maps ​

• design parameter count -> nInput
• objective count -> nObj
• constraint count -> nCon
• bounds -> lb and ub
• variable types -> varType and varSet

How evaluation flows ​

The adapter delegates to SimModel internally. In practice, the flow is:

UQPyL algorithm
  -> Problem.evaluate(X)
  -> UQPyLAdapter
  -> SimModel.run(X)
  -> HydroPilot runtime
  -> objective / constraint arrays

This keeps concerns separate:

• HydroPilot owns model execution, parameter writing, output extraction, and metric evaluation
• UQPyL owns optimization, search policy, and higher-level algorithm behavior

Basic usage ​

from UQPyL.optimization.soea import GA
from hydropilot.integrations import UQPyLAdapter

with UQPyLAdapter("config.yaml") as problem:
    algorithm = GA()
    result = algorithm.run(problem, seed=42)

The adapter can also be called directly through evaluate(X), objFunc(X), and conFunc(X) when you want lower-level control.

When to use it ​

Use this path when:

• the model execution layer is easier to express in HydroPilot
• the search or analysis algorithm is easier to express in UQPyL
• you want to separate hydrological runtime concerns from optimization concerns

Scope and limits ​

UQPyLAdapter is a bridge, not a second analysis layer inside HydroPilot.

What it gives you:

• a UQPyL-compatible Problem
• standard objective and constraint accessors
• reuse of HydroPilot's runtime and config system

What it does not give you:

• UQPyL algorithms themselves
• optimizer-specific convergence guarantees
• HydroPilot-native implementations of UQPyL analysis modules

Read next ​

• Python API
• Examples