Python: Modify config

This covers how the top-level configuration file can be modified using the config keyword argument, directly when running the model.

Consider a rolling optimization use-case, where the model is run multiple times — each time a different offset across the time horizon is used. This may be the base for a model-predictive-control approach or a myopic foresight model. The top-level configuration file could look like this:

config.iesopt.yaml

parameters: global

config:
  general:
    version: 2.7.0
  optimization:
    problem_type: LP
    snapshots:
      offset: 0
      count: 24

carriers:
  # ... all the carriers ...

components:
  # ... all the components ...

Now, let’s assume you want to run that model in a for loop, where the offset is changed each time. You can simply pass the new value of offset to the config keyword argument, like this:

Rolling optimization using Python.

import iesopt

for day in range(0, 365):
    model = iesopt.run(
        "config.iesopt.yaml",
        config = {
            "optimization.snapshots.offset": day * 24,
        }
    )

    # ... do something with the results here ...

As you can see, you can access any entry in the configuration file using a “dot notation”.

Modifying this to feature a longer look-ahead horizon (than the step size) could then also be done from Python, by simply adding a second entry to the config dictionary:

Rolling optimization with a one-week window.

import iesopt

for day in range(0, 365):
    model = iesopt.run(
        "config.iesopt.yaml",
        config = {
            "optimization.snapshots.offset": day * 24,
            "optimization.snapshots.count": 168,
        }
    )

    # ... do something with the results here ...

Tip

You can use this to easily control which input data files are loaded — without the need to manually introduce a parameter that controls each file’s name. Just use the config keyword argument to set, for example, that the file registered as data (see example config) points to a specific weather year’s file, by passing config = {"files.data": "weather_res_2023.csv"}.