rtac.wrapper.tsp

This module implements the target algorithm wrapper for CaDiCaL 1.2.1.

Classes

TSP_Q()

Python-TSP Wrapper for cost minimization scenario.

TSP_Qpp()

Python-TSP Wrapper for cost minimization scenario.

TSP_RT()

Python-TSP Wrapper for runtime minimization scenario.

TSP_RTpp()

Python-TSP Wrapper for runtime minimization scenario.
class rtac.wrapper.tsp.TSP_Q[source]

Bases: TSP_RT

Python-TSP Wrapper for cost minimization scenario. Annealing factor is not fixed. If TA is much faster than the time limit but still yields a better solution it is not a problem.

translate_config(config: Configuration) → list[str][source]

Overriding TSP_RT function: Convert dictionary representation of the configuration to a list of parameter name and value alternating.

Parameters:

config (Configuration) – Configuration object – parameter values to run the problem instance with.

Returns:

List representation of the configuration with parameter names and values alternating.

Return type:

list of str

class rtac.wrapper.tsp.TSP_Qpp[source]

Bases: TSP_Q

Python-TSP Wrapper for cost minimization scenario. Annealing factor is not fixed. If TA is much faster than the time limit but still yields a better solution it is not a problem. Additional functions for ReACTR++ implementation.

check_output(ta_output) → list[float] | None[source]

Parse runtime output of the target algorithm.

Parameters:

ta_output (bytes) – Output of the target algorithm.

Returns:

List of intermediate output values if provided by the target algorithm.

Return type:

list of float or None

interim_info() → list[rtac.ac_functionalities.rtac_data.InterimMeaning][source]

Gives information about whether a higher or a lower level of the entry is a sign of higher quality of the configuration regarding the target algorithm run.

Returns:

Indicates if a higher or lower value is better.

Return type:

list of InterimMeaning or None

class rtac.wrapper.tsp.TSP_RT[source]

Bases: AbstractWrapper

Python-TSP Wrapper for runtime minimization scenario. Annealing factor ‘a’ is fixed to have a fair comparison of runtime performance.

check_if_solved(ta_output: bytes, nnr: non_block_read, proc: Popen) → tuple[int | float, float, int] | None[source]

Check the bytes output of the subprocess.Popen process running CaDiCaL to determine if the problem instance is solved.

Parameters:

  • ta_output (bytes) – Output of the target algorithm.

  • nnr (non_nlock_read) – Non-blocking read function for accessing the subprocess.PIPE output.

  • proc (subprocess.Popen) – Target algorithm run via subprocess.Popen process.

Returns:

Target algorithm result (1 if solved, 0 otherwise), runtime needed, and event (0 or 1, indicating if solved), or None if no result.

Return type:

tuple of (int or float, float, int) or None

start(config: Any, timeout: int, instance: str) → tuple[subprocess.Popen, int][source]

Start CaDiCaL via subprocess.Popen with stdout set to subprocess.PIPE, using the given configuration on the specified instance with a time limit.

Parameters:

  • config (Any) – Parameters in the format required by the target algorithm.

  • timeout (int) – Maximum runtime allowed for the target algorithm run in seconds.

  • instance (str) – Path to the problem instance.

Returns:

Target algorithm subprocess.Popen process and the start time of the process.

Return type:

tuple of (subprocess.Popen, int)

translate_config(config: Configuration) → list[str][source]

Convert dictionary representation of the configuration to a list of parameter names and values alternating.

Parameters:

config (Configuration) – Configuration object - parameter values to run problem instance with.

Returns:

List of strings representation of the configuration.

Return type:

list of str

class rtac.wrapper.tsp.TSP_RTpp[source]

Bases: TSP_RT

Python-TSP Wrapper for runtime minimization scenario. Annealing factor ‘a’ is fixed to have a fair comparison of runtime performance. Additional functions for ReACTR++ implementation.

check_output(ta_output: bytes) → list[float] | None[source]

Parse runtime output of the target algorithm.

Parameters:

ta_output (bytes) – Output of the target algorithm.

Returns:

List of intermediate output values if provided by the target algorithm.

Return type:

list of float or None

interim_info() → list[rtac.ac_functionalities.rtac_data.InterimMeaning][source]

Gives information about whether a higher or a lower level of the entry is a sign of higher quality of the configuration regarding the target algorithm run.

Returns:

Indicates if a higher or lower value is better.

Return type:

list of InterimMeaning or None