pympcc.MPCCResult¶

class pympcc.MPCCResult(x, obj, status, message, G, H, comp_residual, comp_residual_mean, success, strategy, history=<factory>, solve_time=None, mult_g=None, stationarity='unknown', kkt_residual=None, cleanup_status=None, cleanup_n_iter=None, cleanup_obj=None, cleanup_active_set=None, cleanup_accepted=None, comp_G_scale=None, comp_H_scale=None, cq=None, cq_active_set_sizes=None, cq_rank_deficit=None, b_stationary=None, b_stationary_witness=None, b_stationary_min_descent=None, per_pair_status=None, mult_comp_G_mpcc=None, mult_comp_H_mpcc=None, tnlp_refined=None, sosc=None, sosc_min_eigenvalue=None, sosc_skipped_reason=None, jac_condition=None, hessian_condition_estimate=None, time_limit_hit=False, merit_cross_check=None, jac_row_norms=None, jac_col_norms=None, degeneracy_report=None, initial_point_stats=None, n_ipopt_iter_total=None, warmstart_savings_iter=None)[source]¶

Bases: object

Result returned by pympcc.solve() or pympcc.MPCCSolver.solve().

Parameters:

x (ndarray)
obj (float)
status (int)
message (str)
G (ndarray)
H (ndarray)
comp_residual (float)
comp_residual_mean (float)
success (bool)
strategy (str)
history (list[IterationInfo])
solve_time (float | None)
mult_g (ndarray | None)
stationarity (str)
kkt_residual (float | None)
cleanup_status (int | None)
cleanup_n_iter (int | None)
cleanup_obj (float | None)
cleanup_active_set (tuple | None)
cleanup_accepted (bool | None)
comp_G_scale (ndarray | None)
comp_H_scale (ndarray | None)
cq (str | None)
cq_active_set_sizes (dict | None)
cq_rank_deficit (int | None)
b_stationary (str | None)
b_stationary_witness (tuple | None)
b_stationary_min_descent (float | None)
per_pair_status (list | None)
mult_comp_G_mpcc (ndarray | None)
mult_comp_H_mpcc (ndarray | None)
tnlp_refined (Any | None)
sosc (bool | None)
sosc_min_eigenvalue (float | None)
sosc_skipped_reason (str | None)
jac_condition (float | None)
hessian_condition_estimate (float | None)
time_limit_hit (bool)
merit_cross_check (dict | None)
jac_row_norms (dict | None)
jac_col_norms (dict | None)
degeneracy_report (dict | None)
initial_point_stats (dict | None)
n_ipopt_iter_total (int | None)
warmstart_savings_iter (int | None)

x¶

Solution vector.

Type:: ndarray

obj¶

Objective value at the solution.

Type:: float

status¶

IPOPT exit code of the final NLP solve. Common values: 0 = Solve_Succeeded, 1 = Solved_To_Acceptable_Level, -1 = Maximum_Iterations_Exceeded, 2 = Infeasible_Problem_Detected.

Type:: int

message¶

Human-readable status from IPOPT.

Type:: str

G¶

G(x*) values at the solution.

Type:: ndarray, shape (n_comp,)

H¶

H(x*) values at the solution.

Type:: ndarray, shape (n_comp,)

comp_residual¶

Complementarity feasibility measure: max_i |G_i(x*) * H_i(x*)|.

Type:: float

comp_residual_mean¶

Mean complementarity residual: mean_i |G_i(x*) * H_i(x*)|.

Type:: float

success¶

True when IPOPT status is 0 (Solve_Succeeded) or 1 (Solved_To_Acceptable_Level).

Type:: bool

strategy¶

Name of the reformulation strategy used.

Type:: str

solve_time¶

Sum of wall-clock seconds spent inside each nlp.solve() call. Excludes strategy setup (NLP construction, sparsity computation), callback invocations, and post-solve stationarity classification. None when not measured (e.g. results constructed programmatically).

Type:: float or None

history¶

Per-iteration diagnostics for iterative strategies (Scholtes, smoothing). Empty for the direct strategy.

Type:: list[IterationInfo]

__init__(x, obj, status, message, G, H, comp_residual, comp_residual_mean, success, strategy, history=<factory>, solve_time=None, mult_g=None, stationarity='unknown', kkt_residual=None, cleanup_status=None, cleanup_n_iter=None, cleanup_obj=None, cleanup_active_set=None, cleanup_accepted=None, comp_G_scale=None, comp_H_scale=None, cq=None, cq_active_set_sizes=None, cq_rank_deficit=None, b_stationary=None, b_stationary_witness=None, b_stationary_min_descent=None, per_pair_status=None, mult_comp_G_mpcc=None, mult_comp_H_mpcc=None, tnlp_refined=None, sosc=None, sosc_min_eigenvalue=None, sosc_skipped_reason=None, jac_condition=None, hessian_condition_estimate=None, time_limit_hit=False, merit_cross_check=None, jac_row_norms=None, jac_col_norms=None, degeneracy_report=None, initial_point_stats=None, n_ipopt_iter_total=None, warmstart_savings_iter=None)¶

Parameters:

x (ndarray)
obj (float)
status (int)
message (str)
G (ndarray)
H (ndarray)
comp_residual (float)
comp_residual_mean (float)
success (bool)
strategy (str)
history (list[IterationInfo])
solve_time (float | None)
mult_g (ndarray | None)
stationarity (str)
kkt_residual (float | None)
cleanup_status (int | None)
cleanup_n_iter (int | None)
cleanup_obj (float | None)
cleanup_active_set (tuple | None)
cleanup_accepted (bool | None)
comp_G_scale (ndarray | None)
comp_H_scale (ndarray | None)
cq (str | None)
cq_active_set_sizes (dict | None)
cq_rank_deficit (int | None)
b_stationary (str | None)
b_stationary_witness (tuple | None)
b_stationary_min_descent (float | None)
per_pair_status (list | None)
mult_comp_G_mpcc (ndarray | None)
mult_comp_H_mpcc (ndarray | None)
tnlp_refined (Any | None)
sosc (bool | None)
sosc_min_eigenvalue (float | None)
sosc_skipped_reason (str | None)
jac_condition (float | None)
hessian_condition_estimate (float | None)
time_limit_hit (bool)
merit_cross_check (dict | None)
jac_row_norms (dict | None)
jac_col_norms (dict | None)
degeneracy_report (dict | None)
initial_point_stats (dict | None)
n_ipopt_iter_total (int | None)
warmstart_savings_iter (int | None)

Return type:

None

Methods

`__init__`(x, obj, status, message, G, H, ...)
`summary`([verbosity])	Return a formatted summary of the result.
`to_dataframe`()	Return a per-complementarity-pair `pandas.DataFrame`.
`to_json`()	Serialize the result to a JSON string.
`unscale_comp_multipliers`(mpcc_mult_G, ...)	Convert scaled-space comp multipliers back to original space.

Attributes

`b_stationary`
`b_stationary_min_descent`
`b_stationary_witness`
`cleanup_accepted`
`cleanup_active_set`
`cleanup_n_iter`
`cleanup_obj`
`cleanup_status`
`comp_G_scale`
`comp_H_scale`
`cq`
`cq_active_set_sizes`
`cq_rank_deficit`
`degeneracy_report`
`hessian_condition_estimate`
`initial_point_stats`
`jac_col_norms`
`jac_condition`
`jac_row_norms`
`kkt_residual`
`merit_cross_check`
`mult_comp_G_mpcc`
`mult_comp_H_mpcc`
`mult_g`
`n_ipopt_iter_total`
`per_pair_status`
`solve_time`
`sosc`
`sosc_min_eigenvalue`
`sosc_skipped_reason`
`stationarity`
`time_limit_hit`
`tnlp_refined`
`warmstart_savings_iter`
`x`
`obj`
`status`
`message`
`G`
`H`
`comp_residual`
`comp_residual_mean`
`success`
`strategy`
`history`

x: ndarray¶

obj: float¶

status: int¶

message: str¶

G: ndarray¶

H: ndarray¶

comp_residual: float¶

comp_residual_mean: float¶

success: bool¶

strategy: str¶

history: list[IterationInfo]¶

solve_time: float | None = None¶

mult_g: ndarray | None = None¶

stationarity: str = 'unknown'¶

kkt_residual: float | None = None¶

cleanup_status: int | None = None¶

cleanup_n_iter: int | None = None¶

cleanup_obj: float | None = None¶

cleanup_active_set: tuple | None = None¶

cleanup_accepted: bool | None = None¶

comp_G_scale: ndarray | None = None¶

comp_H_scale: ndarray | None = None¶

cq: str | None = None¶

cq_active_set_sizes: dict | None = None¶

cq_rank_deficit: int | None = None¶

b_stationary: str | None = None¶

b_stationary_witness: tuple | None = None¶

b_stationary_min_descent: float | None = None¶

per_pair_status: list | None = None¶

mult_comp_G_mpcc: ndarray | None = None¶

mult_comp_H_mpcc: ndarray | None = None¶

tnlp_refined: Any | None = None¶

sosc: bool | None = None¶

sosc_min_eigenvalue: float | None = None¶

sosc_skipped_reason: str | None = None¶

jac_condition: float | None = None¶

hessian_condition_estimate: float | None = None¶

time_limit_hit: bool = False¶

merit_cross_check: dict | None = None¶

jac_row_norms: dict | None = None¶

jac_col_norms: dict | None = None¶

degeneracy_report: dict | None = None¶

initial_point_stats: dict | None = None¶

n_ipopt_iter_total: int | None = None¶

warmstart_savings_iter: int | None = None¶

unscale_comp_multipliers(mpcc_mult_G, mpcc_mult_H)[source]¶

Convert scaled-space comp multipliers back to original space.

If the problem was solved with comp_G_scale/comp_H_scale active, IPOPT’s multipliers correspond to the scaled constraint s · G(x) ≥ 0. Original-space duals satisfy μ = s · μ̃ (chain rule). Returns the input unchanged if no scaling was used.

Return type:

tuple[ndarray, ndarray]

Parameters:

mpcc_mult_G (ndarray)
mpcc_mult_H (ndarray)

to_json()[source]¶

Serialize the result to a JSON string.

Arrays are converted to lists. Fields that are None are included as JSON null. The history list is omitted (use summary(verbosity=2) for human-readable history).

Returns:: JSON-encoded result suitable for logging, benchmarking, or storage.
Return type:: str

to_dataframe()[source]¶

Return a per-complementarity-pair pandas.DataFrame.

Columns: pair, G, H, GH, status, and (when available from TNLP refinement) mu_G, mu_H.

Requires pandas.

Returns:: One row per complementarity pair.
Return type:: pandas.DataFrame
Raises:: ImportError – When pandas is not installed.

summary(verbosity=1)[source]¶

Return a formatted summary of the result.

Parameters:: verbosity (int) – 0 — single-line headline. 1 — multi-section block (default); shows objective, residuals, stationarity, populated diagnostics, and performance. 2 — also includes the per-iteration history table.
Return type:: str