Exact solver

Solver

This Solver solves certain scheduling problems optimally

It can handle:

  • Classic JSSP

  • Classic FJSSP

  • both of the above with and without tool constraints

  • optimization criteria tardiness and makespan

class agents.solver.solver.OrToolSolver

Bases: object

This class can be used to solve JSSP problems. It can handle:

  • Classic JSSP

  • Classic FJSSP

  • both of the above with and without tool constraints

  • optimization criteria tardiness and makespan

Data needs to be passed in ‘instance format’ and is returned in this format, too.

classmethod optimize(instance: List[Task], objective: str = 'makespan')

Optimizes the passed instance according to the passed objective.

Parameters

  • instance (List[Task]) – The instance as a list of Tasks

  • objective (str) – Bbjective to be minimized. May be ‘makespan’ or ‘tardiness’.

Returns

tuple(list[Task], float) Solved instance and objective value

static parse_instance_to_solver_format(instance: List[Task])

Parses the instance to a processable format.

Parameters

instance (list[Task]) – The instance as a list of Tasks

Returns

jobs lists with Tuples for every task, including all necessary information for the solver function machine_id(s), processing_time, due_date(set to 0 for all but last task in job), tool

Example

Job2 = [([0, 2, 3], 8, 0, 4), ([0, 2, 3], 6, 0, 2), …]

static parse_to_plottable_format(original_instance: List[Task], assigned_jobs_by_solver)

Reformats the solution into the original instance format to be passed to the gantt chart generator.

Parameters

  • original_instance (list[Tasks]) – Original instance in original format

  • assigned_jobs_by_solver – solution passed by the optimize() function

Returns

list[Tasks] plottable solved instance

agents.solver.solver.get_perser_args()

parse arguments from command line

agents.solver.solver.main(instances_data_file_path, solver_objective, write_to_file=False, plot_gantt_chart=False)