discrete_optimization.rcpsp.transformations package

Submodules

discrete_optimization.rcpsp.transformations.generic_scheduling_impl module

Transformation from/to generic scheduling to/from RCPSP.

class discrete_optimization.rcpsp.transformations.generic_scheduling_impl.GenericSchedulingToRcpspTransformation

Bases: FromGenericSchedulingImpl[RcpspProblem, RcpspSolution]

Transform GenericSchedulingImplProblem to RCPSP.

get_forward_metadata() → TransformationMetadata

Get metadata for problem transformation (P1 → P2).

Documents what information is lost when transforming the PROBLEM.

Override this method to provide detailed information about: - Constraints that cannot be represented in target problem - Objectives that are ignored or approximated - Assumptions made during transformation

Returns:

TransformationMetadata documenting losses in problem transformation

Default:

Returns exact_transformation() (no losses documented)

Note

Solutions from the target problem always map back MECHANICALLY via back_transform_solution(), but may not satisfy all constraints from the original source problem if this transformation is lossy.

Example: BinPack → SALBP loses incompatibility constraints.

Solutions from SALBP solvers map back to BinPack allocations, but may violate incompatibility if that constraint was present.

Always verify solutions in the original problem after solving via transformation!

transform_problem(source_problem: GenericSchedulingImplProblem) → RcpspProblem

Parameters:

source_problem

Returns:

class discrete_optimization.rcpsp.transformations.generic_scheduling_impl.RcpspToGenericSchedulingTransformation

Bases: ToGenericSchedulingImpl[RcpspProblem, RcpspSolution]

Transform RCPSP to GenericSchedulingImplProblem.

get_forward_metadata() → TransformationMetadata

Get metadata for problem transformation (P1 → P2).

Documents what information is lost when transforming the PROBLEM.

Override this method to provide detailed information about: - Constraints that cannot be represented in target problem - Objectives that are ignored or approximated - Assumptions made during transformation

Returns:

TransformationMetadata documenting losses in problem transformation

Default:

Returns exact_transformation() (no losses documented)

Note

Solutions from the target problem always map back MECHANICALLY via back_transform_solution(), but may not satisfy all constraints from the original source problem if this transformation is lossy.

Example: BinPack → SALBP loses incompatibility constraints.

Solutions from SALBP solvers map back to BinPack allocations, but may violate incompatibility if that constraint was present.

Always verify solutions in the original problem after solving via transformation!

transform_solution_from_raw_generic_to_specific(raw_sol: RawSolution[Hashable, Hashable, Hashable], source_problem: RcpspProblem) → RcpspSolution

Convert a raw solution (from generic problem) into a specific solution to the source problem.

Parameters:

source_problem

Returns:

discrete_optimization.rcpsp.transformations.generic_scheduling_impl.transform_solution_from_raw_generic_to_rcpsp(raw_sol: RawSolution[Hashable, Hashable, Hashable], problem: RcpspProblem) → RcpspSolution

Convert generic solution to RCPSP solution.

Parameters:

• solution

• problem

Returns:

discrete_optimization.rcpsp.transformations.to_multiskill module

Transformation from RCPSP to RCPSP Multiskill.

class discrete_optimization.rcpsp.transformations.to_multiskill.RcpspToMultiskillTransformation

Bases: ProblemTransformation[RcpspProblem, RcpspSolution, MultiskillRcpspProblem, MultiskillRcpspSolution]

Transform RCPSP to RCPSPMultiskill.

Mapping strategy: - RCPSP resources map directly to MultiskillRCPSP cumulative resources - No employees or skills needed (MultiskillRCPSP can work with just resources) - Same mode details and precedence constraints

Example

# >>> problem = RcpspProblem( # … resources={“R1”: 5, “R2”: 2}, # … … # … ) # >>> transformation = RcpspToMultiskillTransformation() # >>> ms_problem = transformation.transform_problem(problem) # >>> # ms_problem uses resources R1, R2 directly (no fake employees)

back_transform_solution(solution: MultiskillRcpspSolution, source_problem: RcpspProblem) → RcpspSolution

Convert Multiskill solution back to RCPSP solution.

We only need the schedule (start times) and modes. Employee assignments are not relevant (empty in this transformation).

Parameters:

• solution – Solution in multiskill problem space

• source_problem – Original RCPSP problem

Returns:

Corresponding RCPSP solution

forward_transform_solution(solution: RcpspSolution, target_problem: MultiskillRcpspProblem) → MultiskillRcpspSolution

Convert RCPSP solution to Multiskill solution.

Since we don’t use employees, this is straightforward.

Parameters:

• solution – Solution in RCPSP problem space

• target_problem – Transformed multiskill problem

Returns:

Corresponding multiskill solution (without employee assignments)

transform_problem(source_problem: RcpspProblem) → MultiskillRcpspProblem

Convert RCPSP to Multiskill RCPSP.

Parameters:

source_problem – Original RCPSP problem

Returns:

Equivalent MultiskillRcpspProblem

discrete_optimization.rcpsp.transformations.to_preemptive module

Transformation from RCPSP to Preemptive RCPSP.

class discrete_optimization.rcpsp.transformations.to_preemptive.RcpspToPreemptiveTransformation

Bases: ProblemTransformation[RcpspProblem, RcpspSolution, PreemptiveRcpspProblem, PreemptiveRcpspSolution]

Transform RCPSP to Preemptive RCPSP.

Mapping: - All tasks → Preemptive tasks (can be interrupted) - All resources and constraints preserved - Same mode details and precedence - All tasks marked as preemptive

RCPSP is a special case of Preemptive RCPSP where tasks cannot be preempted. The reverse is also true when solutions don’t use preemption.

back_transform_solution(solution: PreemptiveRcpspSolution, source_problem: RcpspProblem) → RcpspSolution

Transform Preemptive RCPSP solution back to RCPSP solution.

Parameters:

• solution – Preemptive RCPSP solution

• source_problem – Original RCPSP problem

Returns:

Equivalent RCPSP solution

Note

If the preemptive solution uses preemption (tasks with multiple parts), the RCPSP solution will use the first start time and last end time, potentially with idle time in between. This may not be feasible in the non-preemptive problem if resources are occupied during idle periods.

forward_transform_solution(solution: RcpspSolution, target_problem: PreemptiveRcpspProblem) → PreemptiveRcpspSolution | None

Transform RCPSP solution to Preemptive RCPSP solution (for warmstart).

Parameters:

• solution – RCPSP solution

• target_problem – Target Preemptive RCPSP problem

Returns:

Equivalent Preemptive RCPSP solution (without preemption)

get_forward_metadata() → TransformationMetadata

Metadata for forward problem transformation (RCPSP → Preemptive RCPSP).

This direction is EXACT: RCPSP is a subset of Preemptive RCPSP.

transform_problem(source_problem: RcpspProblem) → PreemptiveRcpspProblem

Transform RCPSP to Preemptive RCPSP.

Parameters:

source_problem – RCPSP problem instance

Returns:

Equivalent Preemptive RCPSP problem

Module contents

Transformations for RCPSP problems.

class discrete_optimization.rcpsp.transformations.RcpspToMultiskillTransformation

Bases: ProblemTransformation[RcpspProblem, RcpspSolution, MultiskillRcpspProblem, MultiskillRcpspSolution]

Transform RCPSP to RCPSPMultiskill.

Mapping strategy: - RCPSP resources map directly to MultiskillRCPSP cumulative resources - No employees or skills needed (MultiskillRCPSP can work with just resources) - Same mode details and precedence constraints

Example

# >>> problem = RcpspProblem( # … resources={“R1”: 5, “R2”: 2}, # … … # … ) # >>> transformation = RcpspToMultiskillTransformation() # >>> ms_problem = transformation.transform_problem(problem) # >>> # ms_problem uses resources R1, R2 directly (no fake employees)

back_transform_solution(solution: MultiskillRcpspSolution, source_problem: RcpspProblem) → RcpspSolution

Convert Multiskill solution back to RCPSP solution.

We only need the schedule (start times) and modes. Employee assignments are not relevant (empty in this transformation).

Parameters:

• solution – Solution in multiskill problem space

• source_problem – Original RCPSP problem

Returns:

Corresponding RCPSP solution

forward_transform_solution(solution: RcpspSolution, target_problem: MultiskillRcpspProblem) → MultiskillRcpspSolution

Convert RCPSP solution to Multiskill solution.

Since we don’t use employees, this is straightforward.

Parameters:

• solution – Solution in RCPSP problem space

• target_problem – Transformed multiskill problem

Returns:

Corresponding multiskill solution (without employee assignments)

transform_problem(source_problem: RcpspProblem) → MultiskillRcpspProblem

Convert RCPSP to Multiskill RCPSP.

Parameters:

source_problem – Original RCPSP problem

Returns:

Equivalent MultiskillRcpspProblem

class discrete_optimization.rcpsp.transformations.RcpspToPreemptiveTransformation

Bases: ProblemTransformation[RcpspProblem, RcpspSolution, PreemptiveRcpspProblem, PreemptiveRcpspSolution]

Transform RCPSP to Preemptive RCPSP.

Mapping: - All tasks → Preemptive tasks (can be interrupted) - All resources and constraints preserved - Same mode details and precedence - All tasks marked as preemptive

RCPSP is a special case of Preemptive RCPSP where tasks cannot be preempted. The reverse is also true when solutions don’t use preemption.

back_transform_solution(solution: PreemptiveRcpspSolution, source_problem: RcpspProblem) → RcpspSolution

Transform Preemptive RCPSP solution back to RCPSP solution.

Parameters:

• solution – Preemptive RCPSP solution

• source_problem – Original RCPSP problem

Returns:

Equivalent RCPSP solution

Note

If the preemptive solution uses preemption (tasks with multiple parts), the RCPSP solution will use the first start time and last end time, potentially with idle time in between. This may not be feasible in the non-preemptive problem if resources are occupied during idle periods.

forward_transform_solution(solution: RcpspSolution, target_problem: PreemptiveRcpspProblem) → PreemptiveRcpspSolution | None

Transform RCPSP solution to Preemptive RCPSP solution (for warmstart).

Parameters:

• solution – RCPSP solution

• target_problem – Target Preemptive RCPSP problem

Returns:

Equivalent Preemptive RCPSP solution (without preemption)

get_forward_metadata() → TransformationMetadata

Metadata for forward problem transformation (RCPSP → Preemptive RCPSP).

This direction is EXACT: RCPSP is a subset of Preemptive RCPSP.

transform_problem(source_problem: RcpspProblem) → PreemptiveRcpspProblem

Transform RCPSP to Preemptive RCPSP.

Parameters:

source_problem – RCPSP problem instance

Returns:

Equivalent Preemptive RCPSP problem