eb-optimization¶
eb-optimization provides optimization, tuning, and search utilities used to transform forecast outputs into actionable decisions within the Electric Barometer ecosystem.
This package focuses on policy application, parameter tuning, and search strategies, rather than model training or metric definition.
Scope¶
This package is responsible for:
- Applying readiness and cost-ratio policies to forecast outputs
- Estimating and tuning cost-ratio and readiness parameters
- Supporting grid-based and kernel-based search strategies
- Providing reusable optimization primitives for downstream systems
It intentionally avoids defining metrics, data contracts, or evaluation workflows.
Contents¶
-
Policies Readiness and cost-ratio policies that govern decision adjustments
-
Search utilities Tools for exploring parameter spaces and optimization surfaces
-
Tuning helpers Utilities for estimating and calibrating policy parameters