One-Tank Incremental Learning

This example identifies a one-step predictor from the same deterministic water-tank simulation as the offline learning variant. The difference is the learning strategy: training observations are delivered in order, one row at a time, to an incremental learner.

Shared System and Data

The water level \(x\) follows

\[ \dot{x} = \frac{b V(t) - a \sqrt{x}}{A}, \]

with tank area \(A = 5\), outflow rate \(a = 0.5\), inflow rate \(b = 2\), initial level \(x(0) = 1\), and

\[ V(t) = \max\left(0, \sin\left(2 \pi t / 10\right)\right). \]

examples/one_tank/system.py defines this model once as a single-location HybridSystem with no transitions. simulate_one_tank() simulates 0 to 25 seconds at 0.1-second intervals and returns the sampled t and h columns used by both examples.

A three-sample SlidingWindow creates h_0, h_1, and h_2. The learner predicts the next level h_2 from h_0 and h_1.

Incremental Workflow

run_incremental.py performs these steps:

1. Simulate the shared one-tank system and construct sliding-window samples.
2. Split the ordered observations into 80 percent training and 20 percent test data without shuffling.
3. Wrap the training partition in a StreamingOfflineEnvironment with batch size 1.
4. Train a River HoeffdingTreeRegressor through RiverLearner and learn_incremental.
5. Evaluate the final model on the fixed holdout data with mean absolute error and mean squared error.

This preserves the incremental-learning intent while using the same simulation and evaluation boundary as the offline example. The script prints the training time and evaluation report.

Run

From the repository root:

uv run --directory ./examples/one_tank python run_incremental.py

To run both one-tank variants:

just examples-one_tank

See Learning Strategies for the distinction between offline and incremental learning.