Scenario Loader


Scenario Loader#

For compatibility with other simulation tools, ASSUME provides a variety of scenario loaders. These are:

  • CSV - File based scenarios (most flexible)

  • AMIRIS - used to create comparative studies

  • OEDS - create scenarios with the Open Energy Data Server

The possibilities and a short usage guide of the different scenario loaders are explained below:


The CSV loader is the default scenario loader for ASSUME. Everything is configured through a config.yaml file, which describes a market design and references the input series of the agents, as well as the bidding strategies used.

It is introduced in this example, where a small simulation is created from scratch.

If you already have an existing csv scenario, you can load it using the ASSUME CLI like:

assume -c tiny -s example_01a --db-uri postgresql://assume:assume@localhost:5432/assume


The AMIRIS loader can be used to run examples configured for usage with the energy market simulation tool AMIRIS by the DLR.

from assume import World
from assume.scenario import load_amiris_async

# To download some amiris examples run:
# git clone amiris-examples
# next to the assume folder
base_path = f"../amiris-examples/Germany2019"

# Read the scenario from this base path
amiris_scenario = read_amiris_yaml(base_path)

# Configure where to write the output
db_uri = "postgresql://assume:assume@localhost:5432/assume"

# Create a simulation world
world = World(database_uri=db_uri)

# Let the loader add everything to the world

# Run the scenario

This makes it possible to compare or validate results from AMIRIS. If you want to adjust the scenario or change bidding strategies, you currently have to adjust the amiris loader accordingly, as it currently does not use reinforcement learning or different bidding strategies at all. It tries to resemble the behavior of AMIRIS in the best way possible. As AMIRIS currently only supports a single market design (with different support mechanisms), the market design can not be adjusted.


The Open-Energy-Data-Server is a tool that facilitates the aggregation of open research data in a way that allows for easy reuse and structured work. It includes data from the Marktstammdatenregister of Germany, ENTSO-E, and weather datasets, making it versatile for modeling different localized scenarios.

Once you have an Open-Energy-Data-Server running, you can query data for various scenarios and interactively compare your simulation results with the actual data recorded by ENTSO-E using Grafana.

The main configuration required for the Open-Energy-Data-Server involves specifying the NUTS areas that should be simulated, as well as a marketdesign. An example configuration of how this can be used is shown here:

# where to write the simulation output to - can also be the oeds
db_uri = "postgresql://assume:assume@localhost:5432/assume"
world = World(database_uri=db_uri)
# adjust to your institute's database server
infra_uri = "postgresql://readonly:readonly@myoeds-server:5432"

# you can also just use ["DE"] for a simulation of germany with single agents per generation technology
nuts_config = ["DE1", "DEA", "DEB", "DEC", "DED", "DEE", "DEF"]

# define a marketdesign which can be used for the simulation
marketdesign = [
        rr.rrule(rr.HOURLY, interval=24, dtstart=start, until=end),
        [MarketProduct(timedelta(hours=1), 24, timedelta(hours=1))],
        additional_fields=["block_id", "link", "exclusive_id"],
# load the dataset from the database
load_oeds(world, "oeds_mastr_simulation", "my_studycase", infra_uri, marketdesign, nuts_config)

# Run the scenario

If there are different