jijzepttools.modeling.replace.problem

jijzepttools.modeling.replace.problem#

Functions#

replace_problem(→ jijmodeling.Problem)

Replace the expressions in the problem with the given replacer.

generate_always_feasible_problem(→ jijmodeling.Problem)

generate always feasible problem from the given problem

generate_phase1_feasible_problem(→ jijmodeling.Problem)

Generate Phase 1 always feasible problem with slack-only objective.

generate_phase1_feasible_problem_forall(...)

Generate Phase 1 always feasible problem with slack-only objective (forall version).

generate_always_feasible_problem_forall(...)

generate always feasible problem from the given problem

Module Contents#

replace_problem(problem: jijmodeling.Problem, replacer: Callable[[jijzepttools.modeling.utils.type.Expr], jijzepttools.modeling.utils.type.Expr], problem_name: str | None = None) → jijmodeling.Problem#

Replace the expressions in the problem with the given replacer.

Parameters:

  • problem (jm.Problem) – problem to be transformed

  • replacer (tp.Callable[[Expr], Expr]) – function to replace the expressions

Returns:

problem with replaced expressions

Return type:

jm.Problem

generate_always_feasible_problem(problem: jijmodeling.Problem, slack_var_1_prefix: str = 'v', slack_var_2_prefix: str = 'w', big_m_prefix: str = 'bigM', bound=1000000, exclude_constraints: List[str] | None = None) → jijmodeling.Problem#

generate always feasible problem from the given problem This function generates a new problem with slack variables for each constraints and penalty term to the objective function. This problem is always feasible owing to the slack variables unless the lower/upper bound of the slack variables is too small.

Parameters:

  • problem (jm.Problem) – problem to be transformed

  • slack_var_1_prefix (str, optional) – prefix of slack variables. Defaults to “v”.

  • slack_var_2_prefix (str, optional) – prefix of slack variables. Defaults to “w”.

  • big_m_prefix (str, optional) – prefix of big M. Defaults to “bigM”.

  • bound (int, optional) – bound for slack variables. Defaults to 1000000.

  • exclude_constraints (tp.Optional[tp.List[str]], optional) – list of constraint names to exclude from slack variables. Defaults to None.

Returns:

always feasible problem

Return type:

jm.Problem

generate_phase1_feasible_problem(problem: jijmodeling.Problem, slack_var_1_prefix: str = 'v', slack_var_2_prefix: str = 'w', big_m_prefix: str = 'bigM', bound=1000000, exclude_constraints: List[str] | None = None) → jijmodeling.Problem#

Generate Phase 1 always feasible problem with slack-only objective.

This function generates a problem where the objective is ONLY the slack penalty term, ignoring the original objective. Used for Phase 1 of two-phase Benders decomposition to find feasibility first.

Objective: minimize bigM * sum(slack_vars) (original objective NOT included)

Parameters:

  • problem (jm.Problem) – problem to be transformed

  • slack_var_1_prefix (str, optional) – prefix of slack variables. Defaults to “v”.

  • slack_var_2_prefix (str, optional) – prefix of slack variables. Defaults to “w”.

  • big_m_prefix (str, optional) – prefix of big M. Defaults to “bigM”.

  • bound (int, optional) – bound for slack variables. Defaults to 1000000.

  • exclude_constraints (tp.Optional[tp.List[str]], optional) – list of constraint names to exclude from slack variables. Defaults to None.

Returns:

Phase 1 always feasible problem with slack-only objective

Return type:

jm.Problem

generate_phase1_feasible_problem_forall(problem: jijmodeling.Problem, slack_var_1_prefix: str = 'v', slack_var_2_prefix: str = 'w', big_m_prefix: str = 'bigM', bound=100000000, exclude_constraints: List[str] | None = None) → jijmodeling.Problem#

Generate Phase 1 always feasible problem with slack-only objective (forall version).

This function generates a problem where the objective is ONLY the slack penalty term, ignoring the original objective. Supports constraints with forall indices. Used for Phase 1 of two-phase Benders decomposition to find feasibility first.

Objective: minimize bigM * sum(slack_vars) (original objective NOT included)

Parameters:

  • problem (jm.Problem) – problem to be transformed

  • slack_var_1_prefix (str, optional) – prefix of slack variables. Defaults to “v”.

  • slack_var_2_prefix (str, optional) – prefix of slack variables. Defaults to “w”.

  • big_m_prefix (str, optional) – prefix of big M. Defaults to “bigM”.

  • bound (int, optional) – bound for slack variables. Defaults to 100000000.

  • exclude_constraints (tp.Optional[tp.List[str]], optional) – list of constraint names to exclude from slack variables. Defaults to None.

Returns:

Phase 1 always feasible problem with slack-only objective

Return type:

jm.Problem

generate_always_feasible_problem_forall(problem: jijmodeling.Problem, slack_var_1_prefix: str = 'v', slack_var_2_prefix: str = 'w', big_m_prefix: str = 'bigM', bound=100000000, exclude_constraints: List[str] | None = None) → jijmodeling.Problem#

generate always feasible problem from the given problem This function generates a new problem with slack variables for each constraints including forall and penalty term to the objective function. This problem is always feasible owing to the slack variables unless the lower/upper bound of the slack variables is too small.

Parameters:

  • problem (jm.Problem) – problem to be transformed

  • slack_var_1_prefix (str, optional) – prefix of slack variables. Defaults to “v”.

  • slack_var_2_prefix (str, optional) – prefix of slack variables. Defaults to “w”.

  • big_m_prefix (str, optional) – prefix of big M. Defaults to “bigM”.

  • bound (int, optional) – bound for slack variables. Defaults to 100000000.

  • exclude_constraints (tp.Optional[tp.List[str]], optional) – list of constraint names to exclude from slack variables. Defaults to None.

Returns:

always feasible problem

Return type:

jm.Problem

Contents