discrete_optimization.rcpsp.solvers package

Submodules

discrete_optimization.rcpsp.solvers.cp_mzn module

discrete_optimization.rcpsp.solvers.cp_mzn.CPMultimodeWithFakeTaskRcpspSolver(problem: RcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction = None, silent_solve_error: bool = False, **kwargs)

Deprecated since version 0.1.

class discrete_optimization.rcpsp.solvers.cp_mzn.CpModesMultimodeRcpspSolver(problem: RcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction | None = None)

Bases: object

init_model(**args)

problem: RcpspProblem

retrieve_solutions(result, parameters_cp: ParametersCp)

solve(parameters_cp: ParametersCp | None = None, time_limit: float | None = 100.0, nr_solutions: int = 1, all_solutions: bool = False, **args)

Solve the CP problem with minizinc

Parameters:

• parameters_cp – parameters specific to CP solvers

• time_limit – the solve process stops after this time limit (in seconds). If None, no time limit is applied.

• nr_solutions – of not all_solutions, the solve stops after finding nr_solutions

• all_solutions – if True, do not stop when reaching nr_solutions

• **args – passed to init_model()

Returns:

class discrete_optimization.rcpsp.solvers.cp_mzn.CpMultimodePreemptiveRcpspSolver(problem: PreemptiveRcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction | None = None, silent_solve_error: bool = True, **kwargs)

Bases: MinizincCpSolver, RcpspSolver

add_hard_special_constraints(partial_solution)

constraint_duration_string_preemptive_i(task, duration, part_id=1, sign: SignEnum = SignEnum.EQUAL) → str

constraint_duration_to_min_duration_preemptive(task, min_duration)

constraint_end_time_string(task, end_time, sign: SignEnum = SignEnum.EQUAL) → str

constraint_minduration_all_tasks()

constraint_objective_equal_makespan(task_sink)

constraint_objective_makespan()

constraint_objective_max_time_set_of_jobs(set_of_jobs)

constraint_ressource_requirement_at_time_t(time, ressource, ressource_number, sign: SignEnum = SignEnum.LEQ)

constraint_start_time_precomputed()

constraint_start_time_string(task, start_time, sign: SignEnum = SignEnum.EQUAL) → str

constraint_start_time_string_preemptive_i(task, start_time, part_id=1, sign: SignEnum = SignEnum.EQUAL) → str

constraint_sum_of_ending_time(set_subtasks: set[Hashable])

constraint_sum_of_starting_time(set_subtasks: set[Hashable])

constraint_task_to_mode(task_id, mode)

init_model(**args)

Instantiate a CP model instance

Afterwards, self.instance should not be None anymore.

problem: PreemptiveRcpspProblem

retrieve_solution(_output_item: str | None = None, **kwargs: Any) → PreemptiveRcpspSolution

Return a d-o solution from the variables computed by minizinc.

Parameters:

• _output_item – string representing the minizinc solver output passed by minizinc to the solution constructor

• **kwargs – keyword arguments passed by minzinc to the solution contructor containing the objective value (key “objective”), and the computed variables as defined in minizinc model.

Returns:

class discrete_optimization.rcpsp.solvers.cp_mzn.CpMultimodeRcpspSolver(problem: RcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction | None = None, silent_solve_error: bool = False, **kwargs)

Bases: MinizincCpSolver, RcpspSolver

add_hard_special_constraints(partial_solution)

constraint_end_time_string(task, end_time, sign: SignEnum = SignEnum.EQUAL) → str

constraint_objective_equal_makespan(task_sink)

constraint_objective_makespan()

constraint_objective_max_time_set_of_jobs(set_of_jobs)

constraint_start_time_string(task, start_time, sign: SignEnum = SignEnum.EQUAL) → str

constraint_sum_of_ending_time(set_subtasks: set[Hashable])

constraint_sum_of_starting_time(set_subtasks: set[Hashable])

constraint_task_to_mode(task_id, mode)

hyperparameters: list[Hyperparameter] = [EnumHyperparameter(name='cp_solver_name', default=<CpSolverName.CHUFFED: 0>, depends_on=None, name_in_kwargs='cp_solver_name')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

init_model(**args)

Instantiate a CP model instance

Afterwards, self.instance should not be None anymore.

problem: RcpspProblem

retrieve_solution(_output_item: str | None = None, **kwargs: Any) → RcpspSolution

Return a d-o solution from the variables computed by minizinc.

Parameters:

• _output_item – string representing the minizinc solver output passed by minizinc to the solution constructor

• **kwargs – keyword arguments passed by minzinc to the solution contructor containing the objective value (key “objective”), and the computed variables as defined in minizinc model.

Returns:

class discrete_optimization.rcpsp.solvers.cp_mzn.CpNoBoolMultimodeRcpspSolver(problem: RcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction | None = None, silent_solve_error: bool = False, **kwargs)

Bases: MinizincCpSolver, RcpspSolver

constraint_end_time_string(task, end_time, sign: SignEnum = SignEnum.EQUAL) → str

constraint_objective_equal_makespan(task_sink)

constraint_objective_makespan()

constraint_objective_max_time_set_of_jobs(set_of_jobs)

constraint_start_time_string(task, start_time, sign: SignEnum = SignEnum.EQUAL) → str

init_model(**args)

Instantiate a CP model instance

Afterwards, self.instance should not be None anymore.

problem: RcpspProblem

retrieve_solution(_output_item: str | None = None, **kwargs: Any) → RcpspSolution

Return a d-o solution from the variables computed by minizinc.

Parameters:

• _output_item – string representing the minizinc solver output passed by minizinc to the solution constructor

• **kwargs – keyword arguments passed by minzinc to the solution contructor containing the objective value (key “objective”), and the computed variables as defined in minizinc model.

Returns:

class discrete_optimization.rcpsp.solvers.cp_mzn.CpPreemptiveRcpspSolver(problem: PreemptiveRcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction | None = None, silent_solve_error: bool = True, **kwargs)

Bases: CpMultimodePreemptiveRcpspSolver

constraint_ressource_requirement_at_time_t(time, ressource, ressource_number, sign: SignEnum = SignEnum.LEQ)

init_model(**args)

Instantiate a CP model instance

Afterwards, self.instance should not be None anymore.

problem: PreemptiveRcpspProblem

retrieve_solution(**kwargs: Any) → PreemptiveRcpspSolution

Return a d-o solution from the variables computed by minizinc.

Parameters:

**kwargs – keyword arguments passed by minzinc to the solution contructor containing the objective value (key “objective”), and the computed variables as defined in minizinc model.

Returns:

class discrete_optimization.rcpsp.solvers.cp_mzn.CpRcpspSolver(problem: RcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction | None = None, silent_solve_error: bool = True, **kwargs)

Bases: MinizincCpSolver, RcpspSolver

add_hard_special_constraints(partial_solution)

constraint_end_time_string(task, end_time, sign: SignEnum = SignEnum.EQUAL) → str

constraint_objective_equal_makespan(task_sink)

constraint_objective_makespan()

constraint_objective_max_time_set_of_jobs(set_of_jobs)

constraint_start_time_string(task, start_time, sign: SignEnum = SignEnum.EQUAL) → str

constraint_sum_of_ending_time(set_subtasks: set[Hashable])

constraint_sum_of_starting_time(set_subtasks: set[Hashable])

get_stats()

hyperparameters: list[Hyperparameter] = [EnumHyperparameter(name='cp_solver_name', default=<CpSolverName.CHUFFED: 0>, depends_on=None, name_in_kwargs='cp_solver_name')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

init_model(**args)

Instantiate a CP model instance

Afterwards, self.instance should not be None anymore.

problem: RcpspProblem

retrieve_solution(_output_item: str | None = None, **kwargs: Any) → RcpspSolution

Return a d-o solution from the variables computed by minizinc.

Parameters:

• _output_item – string representing the minizinc solver output passed by minizinc to the solution constructor

• **kwargs – keyword arguments passed by minzinc to the solution contructor containing the objective value (key “objective”), and the computed variables as defined in minizinc model.

Returns:

discrete_optimization.rcpsp.solvers.cp_mzn.add_constraints_string(child_instance, list_of_strings)

discrete_optimization.rcpsp.solvers.cp_mzn.add_fake_task_cp_data(rcpsp_problem: RcpspProblem | PreemptiveRcpspProblem, ignore_fake_task: bool = True, max_time_to_consider: int | None = None)

discrete_optimization.rcpsp.solvers.cp_mzn.add_hard_special_constraints(partial_solution: PartialSolution, cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver | CpMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.add_hard_special_constraints_mrcpsp(partial_solution: PartialSolution, cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.add_soft_special_constraints(partial_solution: PartialSolution, cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.add_soft_special_constraints_mrcpsp(partial_solution: PartialSolution, cp_solver: CpMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.define_second_part_objective(weights, name_penalty, equal=False)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_end_window(end_times_window: dict[Hashable, tuple[int, int]], cp_solver: CpRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpPreemptiveRcpspSolver | CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_after_nunit(list_start_after_nunit: list[tuple[Hashable, Hashable, int]], cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_after_nunit_mrcpsp(list_start_after_nunit: list[tuple[Hashable, Hashable, int]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_at_end(list_start_at_end: list[tuple[Hashable, Hashable]], cp_solver: CpRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpPreemptiveRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_at_end_mrcpsp(list_start_at_end: list[tuple[Hashable, Hashable]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_at_end_plus_offset(list_start_at_end_plus_offset, cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_at_end_plus_offset_mrcpsp(list_start_at_end_plus_offset, cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_times(dict_start_times: dict[Hashable, int], cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_times_mrcpsp(dict_start_times: dict[Hashable, int], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_together(list_start_together: list[tuple[Hashable, Hashable]], cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_together_mrcpsp(list_start_together: list[tuple[Hashable, Hashable]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.hard_start_window(start_times_window: dict[Hashable, tuple[int, int]], cp_solver: CpRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpPreemptiveRcpspSolver | CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.precompute_possible_starting_time_interval(problem: PreemptiveRcpspProblem)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_end_window(end_times_window: dict[Hashable, tuple[int, int]], cp_solver: CpRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpPreemptiveRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_end_window_mrcpsp(end_times_window: dict[Hashable, tuple[int, int]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_after_nunit(list_start_after_nunit: list[tuple[Hashable, Hashable, int]], cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_after_nunit_mrcpsp(list_start_after_nunit: list[tuple[Hashable, Hashable, int]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_at_end(list_start_at_end: list[tuple[Hashable, Hashable]], cp_solver: CpRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpPreemptiveRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_at_end_mrcpsp(list_start_at_end: list[tuple[Hashable, Hashable]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_at_end_plus_offset(list_start_at_end_plus_offset: list[tuple[Hashable, Hashable, int]], cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_at_end_plus_offset_mrcpsp(list_start_at_end_plus_offset: list[tuple[Hashable, Hashable, int]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_times(dict_start_times: dict[Hashable, int], cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_times_mrcpsp(dict_start_times: dict[Hashable, int], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_together(list_start_together: list[tuple[Hashable, Hashable]], cp_solver: CpPreemptiveRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_together_mrcpsp(list_start_together: list[tuple[Hashable, Hashable]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_window(start_times_window: dict[Hashable, tuple[int, int]], cp_solver: CpRcpspSolver | CpMultimodePreemptiveRcpspSolver | CpPreemptiveRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn.soft_start_window_mrcpsp(start_times_window: dict[Hashable, tuple[int, int]], cp_solver: CpMultimodeRcpspSolver | CPMultimodeWithFakeTaskRcpspSolver | CpNoBoolMultimodeRcpspSolver)

discrete_optimization.rcpsp.solvers.cp_mzn_models module

class discrete_optimization.rcpsp.solvers.cp_mzn_models.CpModelEnum(value)

Bases: Enum

An enumeration.

MODES = 'modes'

MULTI = 'multi'

MULTI_CALENDAR = 'multi-calendar'

MULTI_CALENDAR_BOXES = 'multi-calendar-boxes'

MULTI_FAKETASKS = 'multi-faketasks'

MULTI_NO_BOOL = 'multi-no-bool'

MULTI_PREEMPTIVE = 'multi-preemptive'

MULTI_PREEMPTIVE_CALENDAR = 'multi-preemptive-calendar'

MULTI_RESOURCE_FEASIBILITY = 'multi-resource-feasibility'

SINGLE = 'single'

SINGLE_PREEMPTIVE = 'single-preemptive'

SINGLE_PREEMPTIVE_CALENDAR = 'single-preemptive-calendar'

discrete_optimization.rcpsp.solvers.cp_mzn_multiscenario module

class discrete_optimization.rcpsp.solvers.cp_mzn_multiscenario.CpMultiscenarioRcpspSolver(problem: AggregRcpspProblem, cp_solver_name: CpSolverName = CpSolverName.CHUFFED, params_objective_function: ParamsObjectiveFunction | None = None, silent_solve_error: bool = False, **kwargs)

Bases: MinizincCpSolver

property base_problem

hyperparameters: list[Hyperparameter] = [EnumHyperparameter(name='cp_solver_name', default=<CpSolverName.CHUFFED: 0>, depends_on=None, name_in_kwargs='cp_solver_name'), CategoricalHyperparameter(name='relax_ordering', default=False, depends_on=None, name_in_kwargs='relax_ordering'), IntegerHyperparameter(name='nb_incoherence_limit', default=3, depends_on=None, name_in_kwargs='nb_incoherence_limit', low=0, high=10, step=1, log=False)]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

init_model(**args)

Instantiate a CP model instance

Afterwards, self.instance should not be None anymore.

problem: AggregRcpspProblem

retrieve_solution(_output_item: str | None = None, **kwargs: Any) → RcpspSolution

Return a d-o solution from the variables computed by minizinc.

Parameters:

• _output_item – string representing the minizinc solver output passed by minizinc to the solution constructor

• **kwargs – keyword arguments passed by minzinc to the solution contructor containing the objective value (key “objective”), and the computed variables as defined in minizinc model.

Returns:

discrete_optimization.rcpsp.solvers.cp_mzn_multiscenario.add_fake_task_cp_data(rcpsp_problem: RcpspProblem | PreemptiveRcpspProblem, ignore_fake_task: bool = True, max_time_to_consider: int | None = None)

discrete_optimization.rcpsp.solvers.cpm module

class discrete_optimization.rcpsp.solvers.cpm.CpmObject(ESD, EFD, LSD, LFD)

Bases: object

set_earliest_finish_date(EFD)

set_earliest_start_date(ESD)

set_latest_finish_date(LFD)

set_latest_start_date(LSD)

class discrete_optimization.rcpsp.solvers.cpm.CpmRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: RcpspSolver

get_first_time_to_do_one_task(resource_avail_in_time, task_id)

problem: RcpspProblem

return_order_cpm()

run_classic_cpm()

run_sgs_on_order(map_nodes: dict[Any, CpmObject], critical_path: list[Any], total_order: list[Any] | None = None, cut_sgs_by_critical=True)

run_sgs_time_loop(map_nodes: dict[Any, CpmObject], critical_path: list[Any], total_order: list[Any] | None = None)

solve(**kwargs) → ResultStorage

Generic solving function.

Parameters:

• callbacks – list of callbacks used to hook into the various stage of the solve

• **kwargs – any argument specific to the solver

Solvers deriving from SolverDo should use callbacks methods .on_step_end(), … during solve(). But some solvers are not yet updated and are just ignoring it.

Returns (ResultStorage): a result object containing potentially a pool of solutions to a discrete-optimization problem

discrete_optimization.rcpsp.solvers.cpm.run_partial_classic_cpm(partial_schedule, cpm_solver)

discrete_optimization.rcpsp.solvers.cpsat module

class discrete_optimization.rcpsp.solvers.cpsat.CpSatCumulativeResourceRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: CpSatRcpspSolver

Specific solver to minimize the minimum resource amount needed to accomplish the scheduling problem. In this version we sum up the resource for each given time to do the resource optimisation.

add_lexico_constraint(obj: str, value: float) → Iterable[Constraint]

Parameters:

• obj – a string representing the desired objective. Should be one of “makespan” or “used_resource”.

• value – the limiting value. If the optimization direction is maximizing, this is a lower bound, else this is an upper bound.

Returns:

the created constraints.

create_cumulative_constraint_and_resource_capa(model: CpModel, resource: str, resource_capacity_var: dict[str, IntVar], interval_var: dict[tuple[Hashable, int], IntervalVar], is_present_var: dict[tuple[Hashable, int], IntVar], fake_task: list[dict[str, int]], use_overlap_for_disjunctive_resource: bool)

create_resource_capacity_var(model: CpModel)

get_lexico_objective_value(obj: str, res: ResultStorage) → float

Get best internal model objective value found by last call to solve().

The default implementation consists in using the fit of the last solution in result_storage. This assumes: - that the last solution is the best one for the objective considered - that no aggregation was performed but rather that the fitness is a TupleFitness

with values in the same order as self.problem.get_objective_names().

Parameters:

• obj – a string representing the desired objective. Should be one of self.get_lexico_objectives_available().

• res – result storage returned by last call to solve().

Returns:

get_lexico_objectives_available() → list[str]

List objectives available for lexico optimization

It corresponds to the labels accepted for obj argument for - set_lexico_objective() - add_lexico_constraint() - get_lexico_objective_value()

Default to self.problem.get_objective_names().

Returns:

hyperparameters: list[Hyperparameter] = [CategoricalHyperparameter(name='use_overlap_for_disjunctive_resource', default=True, depends_on=None, name_in_kwargs='use_overlap_for_disjunctive_resource')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

static implements_lexico_api() → bool

Tell whether this solver is implementing the api for lexicographic optimization.

Should return True only if

• set_lexico_objective()

• add_lexico_constraint()

• get_lexico_objective_value()

have been really implemented, i.e. - calling set_lexico_objective() and add_lexico_constraint()

should actually change the next call to solve(),

• get_lexico_objective_value() should correspond to the internal model objective

init_model(**kwargs)

Init CP model.

retrieve_solution(cpsolvercb: CpSolverSolutionCallback) → RcpspSolution

Construct a do solution from the cpsat solver internal solution.

It will be called each time the cpsat solver find a new solution. At that point, value of internal variables are accessible via cpsolvercb.Value(VARIABLE_NAME).

Parameters:

cpsolvercb – the ortools callback called when the cpsat solver finds a new solution.

Returns:

the intermediate solution, at do format.

set_lexico_objective(obj: str) → None

Update internal model objective.

Parameters:

obj – a string representing the desired objective. Should be one of “makespan” or “used_resource”.

Returns:

class discrete_optimization.rcpsp.solvers.cpsat.CpSatRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: OrtoolsCpSatSolver, RcpspSolver, WarmstartMixin

add_classical_precedence_constraints(model: CpModel, starts_var: dict[Hashable, IntVar], ends_var: dict[Hashable, IntVar])

add_one_mode_selected_per_task(model: CpModel, is_present_var: dict[tuple[Hashable, int], IntVar], interval_per_tasks: dict[Hashable, set[tuple[Hashable, int]]])

create_cumulative_constraint(model: CpModel, resource: str, interval_var: dict[tuple[Hashable, int], IntervalVar], is_present_var: dict[tuple[Hashable, int], IntVar], fake_task: list[dict[str, int]])

create_fake_tasks() → list[dict[str, int]]

Create tasks representing the variable resource availability. :return:

create_mode_pair_constraint(model: CpModel, interval_per_tasks: dict[Hashable, set[tuple[Hashable, int]]], is_present_var: dict[tuple[Hashable, int], IntVar], pair_mode_constraint: PairModeConstraint)

init_model(**kwargs)

Init CP model.

init_temporal_variable(model: CpModel) → tuple[dict[Hashable, IntVar], dict[Hashable, IntVar], dict[tuple[Hashable, int], IntVar], dict[tuple[Hashable, int], IntervalVar], dict[Hashable, set[tuple[Hashable, int]]]]

retrieve_solution(cpsolvercb: CpSolverSolutionCallback) → RcpspSolution

Construct a do solution from the cpsat solver internal solution.

It will be called each time the cpsat solver find a new solution. At that point, value of internal variables are accessible via cpsolvercb.Value(VARIABLE_NAME).

Parameters:

cpsolvercb – the ortools callback called when the cpsat solver finds a new solution.

Returns:

the intermediate solution, at do format.

set_warm_start(solution: RcpspSolution) → None

Make the solver warm start from the given solution.

class discrete_optimization.rcpsp.solvers.cpsat.CpSatResourceRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: CpSatRcpspSolver

Specific solver to minimize the minimum resource amount needed to accomplish the scheduling problem. In this version we don’t sum up the resource at a given time, and it suits/makes sense mostly for disjunctive resource (machines)

add_lexico_constraint(obj: str, value: float) → Iterable[Constraint]

Parameters:

• obj – a string representing the desired objective. Should be one of self.problem.get_objective_names().

• value – the limiting value. If the optimization direction is maximizing, this is a lower bound, else this is an upper bound.

Returns:

the created constraints.

create_cumulative_constraint_and_used_resource(model: CpModel, resource: str, is_used_resource: dict[str, IntVar], interval_var: dict[tuple[Hashable, int], IntervalVar], is_present_var: dict[tuple[Hashable, int], IntVar], fake_task: list[dict[str, int]])

get_lexico_objective_value(obj: str, res: ResultStorage) → float

Get best internal model objective value found by last call to solve().

The default implementation consists in using the fit of the last solution in result_storage. This assumes: - that the last solution is the best one for the objective considered - that no aggregation was performed but rather that the fitness is a TupleFitness

with values in the same order as self.problem.get_objective_names().

Parameters:

• obj – a string representing the desired objective. Should be one of self.get_lexico_objectives_available().

• res – result storage returned by last call to solve().

Returns:

get_lexico_objectives_available() → list[str]

List objectives available for lexico optimization

It corresponds to the labels accepted for obj argument for - set_lexico_objective() - add_lexico_constraint() - get_lexico_objective_value()

Default to self.problem.get_objective_names().

Returns:

static implements_lexico_api() → bool

Tell whether this solver is implementing the api for lexicographic optimization.

Should return True only if

• set_lexico_objective()

• add_lexico_constraint()

• get_lexico_objective_value()

have been really implemented, i.e. - calling set_lexico_objective() and add_lexico_constraint()

should actually change the next call to solve(),

• get_lexico_objective_value() should correspond to the internal model objective

init_model(**kwargs)

Init CP model.

retrieve_solution(cpsolvercb: CpSolverSolutionCallback) → RcpspSolution

Construct a do solution from the cpsat solver internal solution.

It will be called each time the cpsat solver find a new solution. At that point, value of internal variables are accessible via cpsolvercb.Value(VARIABLE_NAME).

Parameters:

cpsolvercb – the ortools callback called when the cpsat solver finds a new solution.

Returns:

the intermediate solution, at do format.

set_lexico_objective(obj: str) → None

Update internal model objective.

Parameters:

obj – a string representing the desired objective. Should be one of “makespan” or “used_resource”.

Returns:

discrete_optimization.rcpsp.solvers.dp module

class discrete_optimization.rcpsp.solvers.dp.DpRcpspModeling(value)

Bases: Enum

An enumeration.

TASK_AND_TIME = 0

TASK_AVAIL_TASK_UPD = 2

TASK_MULTIMODE = 3

TASK_ORIGINAL = 1

class discrete_optimization.rcpsp.solvers.dp.DpRcpspSolver(problem: RcpspProblem, **kwargs: Any)

Bases: DpSolver, RcpspSolver, WarmstartMixin

hyperparameters: list[Hyperparameter] = [CategoricalHyperparameter(name='solver', default=<class 'builtins.CABS'>, depends_on=None, name_in_kwargs='solver'), EnumHyperparameter(name='modeling', default=<DpRcpspModeling.TASK_MULTIMODE: 3>, depends_on=None, name_in_kwargs='modeling'), CategoricalHyperparameter(name='dual_bound', default=False, depends_on=None, name_in_kwargs='dual_bound')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

init_model(**kwargs: Any)

Initialize internal model used to solve.

Can initialize a ortools, milp, gurobi, … model.

init_model_multimode(**kwargs: Any) → None

init_model_multimode_calendar(**kwargs: Any) → None

init_model_task(**kwargs: Any) → None

init_model_task_time(**kwargs: Any) → None

init_model_task_time_calendar(**kwargs: Any) → None

init_model_task_upd(**kwargs: Any) → None

modeling: DpRcpspModeling

problem: RcpspProblem

retrieve_solution(sol: Solution) → Solution

retrieve_solution_multimode(sol: Solution) → Solution

retrieve_solution_task(sol: Solution) → Solution

retrieve_solution_task_time(sol: Solution) → Solution

set_warm_start(solution: RcpspSolution) → None

Make the solver warm start from the given solution.

transitions: dict

discrete_optimization.rcpsp.solvers.dp.create_resource_consumption_from_calendar(calendar_availability: ndarray) → list[dict[str, int]]

discrete_optimization.rcpsp.solvers.ga module

class discrete_optimization.rcpsp.solvers.ga.GaMultimodeRcpspSolver(problem: Problem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs: Any)

Bases: RcpspSolver

hyperparameters: list[Hyperparameter] = [EnumHyperparameter(name='crossover', default=None, depends_on=None, name_in_kwargs='crossover'), EnumHyperparameter(name='selection', default=<DeapSelection.SEL_TOURNAMENT: 0>, depends_on=None, name_in_kwargs='selection'), IntegerHyperparameter(name='pop_size', default=100, depends_on=None, name_in_kwargs='pop_size', low=1, high=1000, step=1, log=False), FloatHyperparameter(name='mut_rate', default=0.1, depends_on=None, name_in_kwargs='mut_rate', low=0, high=0.9, suggest_low=False, suggest_high=False, step=None, log=False), FloatHyperparameter(name='crossover_rate', default=0.9, depends_on=None, name_in_kwargs='crossover_rate', low=0, high=1, suggest_low=False, suggest_high=False, step=None, log=False), FloatHyperparameter(name='tournament_size', default=0.2, depends_on=None, name_in_kwargs='tournament_size', low=0, high=1, suggest_low=False, suggest_high=False, step=None, log=False)]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

problem: RcpspProblem

solve(parameters_ga: ~discrete_optimization.generic_tools.ea.ga_tools.ParametersAltGa = <discrete_optimization.generic_tools.ea.ga_tools.ParametersAltGa object>, **args)

Generic solving function.

Parameters:

• callbacks – list of callbacks used to hook into the various stage of the solve

• **kwargs – any argument specific to the solver

Solvers deriving from SolverDo should use callbacks methods .on_step_end(), … during solve(). But some solvers are not yet updated and are just ignoring it.

Returns (ResultStorage): a result object containing potentially a pool of solutions to a discrete-optimization problem

class discrete_optimization.rcpsp.solvers.ga.GaRcpspSolver(problem: Problem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs: Any)

Bases: RcpspSolver

hyperparameters: list[Hyperparameter] = [EnumHyperparameter(name='crossover', default=None, depends_on=None, name_in_kwargs='crossover'), EnumHyperparameter(name='selection', default=<DeapSelection.SEL_TOURNAMENT: 0>, depends_on=None, name_in_kwargs='selection'), IntegerHyperparameter(name='pop_size', default=100, depends_on=None, name_in_kwargs='pop_size', low=1, high=1000, step=1, log=False), FloatHyperparameter(name='mut_rate', default=0.1, depends_on=None, name_in_kwargs='mut_rate', low=0, high=0.9, suggest_low=False, suggest_high=False, step=None, log=False), FloatHyperparameter(name='crossover_rate', default=0.9, depends_on=None, name_in_kwargs='crossover_rate', low=0, high=1, suggest_low=False, suggest_high=False, step=None, log=False), FloatHyperparameter(name='tournament_size', default=0.2, depends_on=None, name_in_kwargs='tournament_size', low=0, high=1, suggest_low=False, suggest_high=False, step=None, log=False)]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

problem: RcpspProblem

solve(parameters_ga: ParametersGa | None = None, **args)

Generic solving function.

Parameters:

• callbacks – list of callbacks used to hook into the various stage of the solve

• **kwargs – any argument specific to the solver

Solvers deriving from SolverDo should use callbacks methods .on_step_end(), … during solve(). But some solvers are not yet updated and are just ignoring it.

Returns (ResultStorage): a result object containing potentially a pool of solutions to a discrete-optimization problem

discrete_optimization.rcpsp.solvers.lns_cp module

class discrete_optimization.rcpsp.solvers.lns_cp.PostProcessLeftShift(rcpsp_problem: RcpspProblem, partial_solution: PartialSolution | None = None)

Bases: PostProcessSolution

build_other_solution(result_storage: ResultStorage) → ResultStorage

discrete_optimization.rcpsp.solvers.lns_cp_preemptive module

class discrete_optimization.rcpsp.solvers.lns_cp_preemptive.PostProLeftShift(problem: PreemptiveRcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, do_ls: bool = False, **kwargs)

Bases: PostProcessSolution

build_other_solution(result_storage: ResultStorage) → ResultStorage

discrete_optimization.rcpsp.solvers.lns_cp_preemptive.last_opti_solution(last_result_store: ResultStorage)

discrete_optimization.rcpsp.solvers.lns_cp_preemptive.sgs_variant(solution: PreemptiveRcpspSolution, problem: PreemptiveRcpspProblem, predecessors_dict)

discrete_optimization.rcpsp.solvers.lns_lp module

class discrete_optimization.rcpsp.solvers.lns_lp.GurobiStartTimeIntervalMultimodeRcpspConstraintHandler(problem: RcpspProblem, fraction_to_fix: float = 0.9, minus_delta: int = 2, plus_delta: int = 2)

Bases: GurobiConstraintHandler, _BaseStartTimeIntervalMultimodeRcpspConstraintHandler

adding_constraint_from_results_store(solver: GurobiMultimodeRcpspSolver, result_storage: ResultStorage, **kwargs: Any) → Iterable[Any]

class discrete_optimization.rcpsp.solvers.lns_lp.InitialRcpspMethod(value)

Bases: Enum

An enumeration.

CP = 5

DUMMY = 0

GA = 4

LS = 3

PILE = 1

PILE_CALENDAR = 2

class discrete_optimization.rcpsp.solvers.lns_lp.InitialRcpspSolution(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, initial_method: InitialRcpspMethod = InitialRcpspMethod.PILE)

Bases: InitialSolution

get_starting_solution() → ResultStorage

class discrete_optimization.rcpsp.solvers.lns_lp.MathOptStartTimeIntervalMultimodeRcpspConstraintHandler(problem: RcpspProblem, fraction_to_fix: float = 0.9, minus_delta: int = 2, plus_delta: int = 2)

Bases: OrtoolsMathOptConstraintHandler, _BaseStartTimeIntervalMultimodeRcpspConstraintHandler

adding_constraint_from_results_store(solver: MathOptMultimodeRcpspSolver, result_storage: ResultStorage, **kwargs: Any) → Iterable[Any]

class discrete_optimization.rcpsp.solvers.lns_lp.MathoptFixStartTimeRcpspConstraintHandler(problem: RcpspProblem, fraction_fix_start_time: float = 0.9)

Bases: OrtoolsMathOptConstraintHandler

adding_constraint_from_results_store(solver: MathOptRcpspSolver, result_storage: ResultStorage, **kwargs: Any) → Iterable[Any]

class discrete_optimization.rcpsp.solvers.lns_lp.MathoptStartTimeIntervalRcpspConstraintHandler(problem: RcpspProblem, fraction_to_fix: float = 0.9, minus_delta: int = 2, plus_delta: int = 2)

Bases: OrtoolsMathOptConstraintHandler

adding_constraint_from_results_store(solver: MathOptRcpspSolver, result_storage: ResultStorage, **kwargs: Any) → Iterable[Any]

discrete_optimization.rcpsp.solvers.lp module

class discrete_optimization.rcpsp.solvers.lp.CplexMultimodeRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: CplexMilpSolver, _BaseLpMultimodeRcpspSolver

hyperparameters: list[Hyperparameter] = [CategoricalHyperparameter(name='greedy_start', default=True, depends_on=None, name_in_kwargs='greedy_start')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

init_model(**args)

Initialize internal model used to solve.

Can initialize a ortools, milp, gurobi, … model.

class discrete_optimization.rcpsp.solvers.lp.GurobiMultimodeRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: GurobiMilpSolver, _BaseLpMultimodeRcpspSolver

convert_to_variable_values(solution: RcpspSolution) → dict[gurobipy.Var, float]

Convert a solution to a mapping between model variables and their values.

Will be used by set_warm_start().

hyperparameters: list[Hyperparameter] = [CategoricalHyperparameter(name='greedy_start', default=True, depends_on=None, name_in_kwargs='greedy_start')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

update_model() → None

Update model (especially for gurobi).

It ensures the well defintion of variables.

class discrete_optimization.rcpsp.solvers.lp.MathOptMultimodeRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: OrtoolsMathOptMilpSolver, _BaseLpMultimodeRcpspSolver

convert_to_variable_values(solution: RcpspSolution) → dict[Variable, float]

Convert a solution to a mapping between model variables and their values.

Will be used by set_warm_start().

hyperparameters: list[Hyperparameter] = [CategoricalHyperparameter(name='greedy_start', default=True, depends_on=None, name_in_kwargs='greedy_start')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

max_horizon: int | None = None

partial_solution: PartialSolution | None = None

class discrete_optimization.rcpsp.solvers.lp.MathOptRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs: Any)

Bases: OrtoolsMathOptMilpSolver, _BaseLpRcpspSolver

convert_to_variable_values(solution: RcpspSolution) → dict[Variable, float]

Convert a solution to a mapping between model variables and their values.

Will be used by set_warm_start() to provide a suitable SolutionHint.variable_values. See https://or-tools.github.io/docs/pdoc/ortools/math_opt/python/model_parameters.html#SolutionHint for more information.

Override it in subclasses to have a proper warm start.

hyperparameters: list[Hyperparameter] = [CategoricalHyperparameter(name='greedy_start', default=True, depends_on=None, name_in_kwargs='greedy_start')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

problem: RcpspProblem

discrete_optimization.rcpsp.solvers.lp_gantt module

class discrete_optimization.rcpsp.solvers.lp_gantt.ConstraintTaskIndividual(list_tuple)

Bases: object

list_tuple: list[tuple[str, int, int, bool]]

class discrete_optimization.rcpsp.solvers.lp_gantt.ConstraintWorkDuration(ressource, individual, time_bounds, working_time_upper_bound)

Bases: object

individual: int

ressource: str

time_bounds: tuple[int, int]

working_time_upper_bound: int

class discrete_optimization.rcpsp.solvers.lp_gantt.GurobiGanttMultimodeRcpspSolver(problem: RcpspProblem, rcpsp_solution: RcpspSolution, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: GurobiMilpSolver, _BaseLpGanttMultimodeRcpspSolver

adding_constraint(constraint_description: ConstraintTaskIndividual | ConstraintWorkDuration, constraint_name: str = '')

build_objective_function_from_a_solution(ressource_usage: dict[str, dict[int, dict[int, bool]]], ignore_tuple: set[tuple[str, int, int]] | None = None)

convert_to_variable_values(solution: Solution) → dict[gurobipy.Var, float]

Convert a solution to a mapping between model variables and their values.

Will be used by set_warm_start().

Override it in subclasses to have a proper warm start. You can also override set_warm_start() if default behaviour is not sufficient.

init_model(**args)

Initialize internal model used to solve.

Can initialize a ortools, milp, gurobi, … model.

retrieve_current_solution(get_var_value_for_current_solution: Callable[[Any], float], get_obj_value_for_current_solution: Callable[[], float]) → tuple[dict[Any, dict[Any, dict[Any, Any]]], float]

Retrieve current solution from internal gurobi solution.

This converts internal gurobi solution into a discrete-optimization Solution. This method can be called after the solve in retrieve_solutions() or during solve within a gurobi/pymilp/cplex callback. The difference will be the get_var_value_for_current_solution and get_obj_value_for_current_solution callables passed.

Parameters:

• get_var_value_for_current_solution – function extracting the value of the given variable for the current solution will be different when inside a callback or after the solve is finished

• get_obj_value_for_current_solution – function extracting the value of the objective for the current solution.

Returns:

the converted solution at d-o format

discrete_optimization.rcpsp.solvers.lp_gantt.intersect(i1, i2)

discrete_optimization.rcpsp.solvers.pile module

class discrete_optimization.rcpsp.solvers.pile.Executor(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: PileRcpspSolver

compute_schedule_from_priority_list(permutation_jobs: list[int], modes_dict: dict[int, int])

update(rcpsp_problem: RcpspProblem)

class discrete_optimization.rcpsp.solvers.pile.GreedyChoice(value)

Bases: Enum

An enumeration.

FASTEST = 3

MOST_SUCCESSORS = 1

SAMPLE_MOST_SUCCESSORS = 2

TOTALLY_RANDOM = 4

class discrete_optimization.rcpsp.solvers.pile.PileCalendarRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: RcpspSolver

hyperparameters: list[Hyperparameter] = [EnumHyperparameter(name='greedy_choice', default=<GreedyChoice.MOST_SUCCESSORS: 1>, depends_on=None, name_in_kwargs='greedy_choice')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

problem: RcpspProblem

solve(**kwargs) → ResultStorage

Generic solving function.

Parameters:

• callbacks – list of callbacks used to hook into the various stage of the solve

• **kwargs – any argument specific to the solver

Solvers deriving from SolverDo should use callbacks methods .on_step_end(), … during solve(). But some solvers are not yet updated and are just ignoring it.

Returns (ResultStorage): a result object containing potentially a pool of solutions to a discrete-optimization problem

class discrete_optimization.rcpsp.solvers.pile.PileRcpspSolver(problem: RcpspProblem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs)

Bases: RcpspSolver

hyperparameters: list[Hyperparameter] = [EnumHyperparameter(name='greedy_choice', default=<GreedyChoice.MOST_SUCCESSORS: 1>, depends_on=None, name_in_kwargs='greedy_choice')]

Hyperparameters available for this solver.

These hyperparameters are to be feed to **kwargs found in

• __init__()

• init_model() (when available)

• solve()

solve(**kwargs) → ResultStorage

Generic solving function.

Parameters:

• callbacks – list of callbacks used to hook into the various stage of the solve

• **kwargs – any argument specific to the solver

Solvers deriving from SolverDo should use callbacks methods .on_step_end(), … during solve(). But some solvers are not yet updated and are just ignoring it.

Returns (ResultStorage): a result object containing potentially a pool of solutions to a discrete-optimization problem

discrete_optimization.rcpsp.solvers.rcpsp_solver module

class discrete_optimization.rcpsp.solvers.rcpsp_solver.RcpspSolver(problem: Problem, params_objective_function: ParamsObjectiveFunction | None = None, **kwargs: Any)

Bases: SolverDO

problem: RcpspProblem | PreemptiveRcpspProblem

Module contents