msp.utils package

Submodules

msp.utils.configs module

The mps.utils.configs module defines load_yaml_flag function.

msp.utils.configs.load_yaml_flag(flag)

msp.utils.metric module

The mps.utils.metric module defines tensorflow based metric to track the model performance.

class msp.utils.metric.MeanMakespan(*args, **kwargs)

Bases: keras.metrics.Metric

reset_states()

Resets all of the metric state variables.

result()

Computes and returns the metric value tensor.

update_state(makespan_train, makespan_baseline)

Accumulates statistics for the metric

Parameters

• makespan_train – makespan values for batch data by train model

• makespan_baseline – makespan values for batch data by baseline model

msp.utils.objective module

The mps.utils.objective module defines objective function of MSP problem.

msp.utils.objective.compute_makespan(inputs, schedules)

Compute makespan for MSP problem.

Module contents

The mps.utils module defines several utilities.

class msp.utils.MSPEnv(**kwargs)

Bases: tensorflow.python.module.module.Module

build(input_shape)

current_time_step()

Returns the current TimeStep.

reset()

Returns the current TimeStep after resetting the Environment.

step(actions)

Applies the action and returns the new TimeStep.

class msp.utils.MSPState

Bases: tensorflow.python.module.module.Module

build(input_shape)

Create variables on first call.

property first_node

get_mask()

get_step_count()

property last_node

reset()

update(inputs, selected_node)