Config Data

Config data provides initialization parameters to solvers and optimizers.
It is written as a Python dictionary, not JSON, in order to:

  • Allow direct use of numpy arrays

  • Enable preprocessing and scripting flexibility

During class instantiation, the config data is automatically loaded.
If no config path is provided, the program attempts to read a default file from:

config/*_config.py

in the current working directory.

1. General Solvers

For all standard solvers (DPBESolver, PBMSolver, MCPBESolver),
the config data is a flat dictionary (single level).

  • Each key corresponds to an attribute of the base solver class.

  • On initialization, all keys are directly mapped to class attributes.

This ensures simple and transparent parameter control.

2. ExtruderPBESolver

Because ExtruderPBESolver involves multiple computational regions,
its config structure is slightly more complex.

  • The top-level dictionary contains two keys:

    • geom_params → parameters for the extruder geometry

    • pbe_params → PBE-related parameters for each region

  • Inside pbe_params:

    • global → shared parameters across all regional DPBE solvers (e.g., grid definition)

    • local_0, local_1, … → region-specific parameters, where local_i defines the parameters for the i-th region

This design allows each region to have its own DPBE configuration while sharing common structural definitions.

3. Optimization Framework

For the optimization framework (OptBase), the config data has a multi-layered structure.

  • multi_flag / single_case

    • Control whether OptBase uses OptCore or OptCoreMulti.

  • algo_params

    • Defines optimization algorithm parameters (population size, iteration limits, etc.).

  • pop_params

    • Provides control parameters for the embedded DPBE solver.

  • opt_params

    • Specifies kernel parameters to be optimized, along with their search ranges.

The nested dictionaries (algo_params, pop_params, opt_params) are passed directly to attributes of OptCore or OptCoreMulti, depending on the optimization mode.