Run a model¶

Five steps from pip install to an answer read back, on the dispatch model of the home page: three generators meet a load over four snapshots at least cost. What a spec may contain is the language's to say.

1. Install¶

pip install specsolve

That brings polars, HiGHS and the language.

2. Save the spec¶

Copy the YAML from the home page into dispatch.yaml. It is also examples/dispatch.yaml in the repository.

3. Check the file¶

import specsolve as sps

program = sps.check('dispatch.yaml')

check raises if the file uses something outside the language, and needs no data to do it.

4. Attach the numbers and solve¶

The file declares three parameters and two dimensions. sources supplies each by name. A parameter over one dimension is a table with that dimension and a value column; a bare sequence supplies a dimension's labels:

import polars as pl

generators = ['wind', 'solar', 'gas']
sources = {
    'p_max': pl.DataFrame({'generator': generators, 'value': [80.0, 0.0, 200.0]}),
    'cost': pl.DataFrame({'generator': generators, 'value': [10.0, 25.0, 50.0]}),
    'load': pl.DataFrame({'snapshot': range(4), 'value': [60.0, 120.0, 180.0, 90.0]}),
    'snapshot': range(4),
    'generator': generators,
}

result = sps.solve('dispatch.yaml', sources)
print(result.objective)  # 10500.0

Wind at 10 runs first, and gas at 50 covers what is left. Solar has no capacity, so the where: "p_max > 0" on p built no column for it.

5. Read the answer back¶

print(result.primal('p'))  # (snapshot, generator, value): eight rows, wind and gas at each snapshot
print(result.dual('power_balance'))  # (snapshot, value): the price of one more unit of load

Each answer is a polars table keyed by the declaration's labels. The dual is the cost of the last generator on: 10 at snapshot 0, where wind alone covers the load, and 50 at the other three.

To hand the model to another tool instead, write it. The suffix picks the format:

sps.write('dispatch.yaml', sources, 'dispatch.lp')

Where next¶