ltbams 0.9.8__tar.gz → 0.9.10__tar.gz

{ltbams-0.9.8/ltbams.egg-info → ltbams-0.9.10}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ltbams
-Version: 0.9.8
+Version: 0.9.10
 Summary: Python software for scheduling modeling and co-simulation with dynanics.
 Home-page: https://github.com/CURENT/ams
 Author: Jinning Wang
@@ -41,7 +41,6 @@ Python Software for Power System Scheduling Modeling and Co-Simulation with Dyna
 ![Repo Size](https://img.shields.io/github/repo-size/CURENT/ams)
 [![GitHub last commit (master)](https://img.shields.io/github/last-commit/CURENT/ams/master?label=last%20commit%20to%20master)](https://github.com/CURENT/ams/commits/master/)
 [![GitHub last commit (develop)](https://img.shields.io/github/last-commit/CURENT/ams/develop?label=last%20commit%20to%20develop)](https://github.com/CURENT/ams/commits/develop/)
-[![HitCount](https://hits.dwyl.com/CURENT/ams.svg)](https://hits.dwyl.com/CURENT/ams)
 [![libraries](https://img.shields.io/librariesio/release/pypi/ltbams)](https://libraries.io/pypi/ltbams)
 [![Structure](https://img.shields.io/badge/code_base-visualize-blue)](https://mango-dune-07a8b7110.1.azurestaticapps.net/?repo=CURENT%2Fams)
 
@@ -92,7 +91,10 @@ Use the following resources to get involved.
 
 # Installation
 
-***NOTE:*** ``kvxopt`` is recommended to install via ``conda`` as sometimes ``pip`` struggles to set the correct path for compiled libraries.
+***NOTE:***
+- Version **0.9.9** has known issues. Please avoid using this version
+- `kvxopt` is recommended to install via `conda` as sometimes ``pip`` struggles to set the correct path for compiled libraries
+- `cvxpy` versions **below 1.5** are incompatible with `numpy` versions **2.0 and above**
 
 AMS is released as ``ltbams`` on PyPI and conda-forge.
 Install from PyPI using pip:

{ltbams-0.9.8 → ltbams-0.9.10}/README.md

@@ -8,7 +8,6 @@ Python Software for Power System Scheduling Modeling and Co-Simulation with Dyna
 ![Repo Size](https://img.shields.io/github/repo-size/CURENT/ams)
 [![GitHub last commit (master)](https://img.shields.io/github/last-commit/CURENT/ams/master?label=last%20commit%20to%20master)](https://github.com/CURENT/ams/commits/master/)
 [![GitHub last commit (develop)](https://img.shields.io/github/last-commit/CURENT/ams/develop?label=last%20commit%20to%20develop)](https://github.com/CURENT/ams/commits/develop/)
-[![HitCount](https://hits.dwyl.com/CURENT/ams.svg)](https://hits.dwyl.com/CURENT/ams)
 [![libraries](https://img.shields.io/librariesio/release/pypi/ltbams)](https://libraries.io/pypi/ltbams)
 [![Structure](https://img.shields.io/badge/code_base-visualize-blue)](https://mango-dune-07a8b7110.1.azurestaticapps.net/?repo=CURENT%2Fams)
 
@@ -59,7 +58,10 @@ Use the following resources to get involved.
 
 # Installation
 
-***NOTE:*** ``kvxopt`` is recommended to install via ``conda`` as sometimes ``pip`` struggles to set the correct path for compiled libraries.
+***NOTE:***
+- Version **0.9.9** has known issues. Please avoid using this version
+- `kvxopt` is recommended to install via `conda` as sometimes ``pip`` struggles to set the correct path for compiled libraries
+- `cvxpy` versions **below 1.5** are incompatible with `numpy` versions **2.0 and above**
 
 AMS is released as ``ltbams`` on PyPI and conda-forge.
 Install from PyPI using pip:

{ltbams-0.9.8 → ltbams-0.9.10}/ams/_version.py

@@ -8,11 +8,11 @@ import json
 
 version_json = '''
 {
- "date": "2024-06-18T14:45:25-0400",
+ "date": "2024-09-03T22:41:38-0400",
  "dirty": false,
  "error": null,
- "full-revisionid": "63b55b4b38c980ba34cd7a2fa97dcffb2e4563d9",
- "version": "0.9.8"
+ "full-revisionid": "8bc9504a8281c16236c73decd8e7365b682c706c",
+ "version": "0.9.10"
 }
 '''  # END VERSION_JSON
 

{ltbams-0.9.8 → ltbams-0.9.10}/ams/cli.py

@@ -111,7 +111,7 @@ def preamble():
 
 def main():
     """
-    Entry point of the ANDES command-line interface.
+    Entry point of the AMS command-line interface.
     """
 
     parser = create_parser()

{ltbams-0.9.8 → ltbams-0.9.10}/ams/core/matprocessor.py

@@ -437,22 +437,37 @@ class MatProcessor:
         return b
 
     def build_ptdf(self, line=None, no_store=False,
-                   incremental=False, step=1000, no_tqdm=False):
+                   incremental=False, step=1000, no_tqdm=False,
+                   permc_spec=None, use_umfpack=True):
         """
-        Build the Power Transfer Distribution Factor (PTDF) matrix and store
-        it in the MParam `PTDF` by default.
+        Build the Power Transfer Distribution Factor (PTDF) matrix and optionally store it in `MParam.PTDF`.
 
-        `PTDF[m, n]` means the increased line flow on line `m` when there is
-        1 p.u. power injection at bus `n`. It is very similar to Generation
-        Shift Factor (GSF).
+        PTDF[m, n] represents the increased line flow on line `m` for a 1 p.u. power injection at bus `n`.
+        It is similar to the Generation Shift Factor (GSF).
 
-        Note that there is discrepency between the PTDF-based line flow and
-        DCOPF calcualted line flow. The gap is ignorable for small cases.
+        Note: There may be minor discrepancies between PTDF-based line flow and DCOPF-calculated line flow.
 
-        It requires DC Bbus and Bf.
+        For large cases, use `incremental=True` to calculate the sparse PTDF in chunks, which will be stored
+        as a `scipy.sparse.lil_matrix`. In this mode, the PTDF is calculated in chunks, and a progress bar
+        will be shown unless `no_tqdm=True`.
 
-        For large cases where memory is a concern, use `incremental=True` to
-        calculate the sparse PTDF in chunks in the format of scipy.sparse.lil_matrix.
+        Parameters
+        ----------
+        line : int, str, list, optional
+            Line indices for which the PTDF is calculated. If specified, the PTDF will not be stored in `MParam`.
+        no_store : bool, optional
+            If False, store the PTDF in `MatProcessor.PTDF._v`. Default is False.
+        incremental : bool, optional
+            If True, calculate the sparse PTDF in chunks to save memory. Default is False.
+        step : int, optional
+            Step size for incremental calculation. Default is 1000.
+        no_tqdm : bool, optional
+            If True, disable the progress bar. Default is False.
+        permc_spec : str, optional
+            Column permutation strategy for sparsity preservation. Default is 'COLAMD'.
+        use_umfpack : bool, optional
+            If True, use UMFPACK as the solver. Effective only when (`incremental=True`)
+            & (`line` contains a single line or `step` is 1). Default is True.
 
         Parameters
         ----------
@@ -468,6 +483,10 @@ class MatProcessor:
             Step for incremental calculation.
         no_tqdm : bool, optional
             If True, the progress bar will be disabled.
+        permc_spec : str, optional
+            How to permute the columns of the matrix for sparsity preservation. (default: 'COLAMD')
+        use_umfpack : bool, optional
+            If True, use UMFPACK as the solver. (default: True)
 
         Returns
         -------
@@ -530,8 +549,10 @@ class MatProcessor:
             # NOTE: for PTDF, we are building rows by rows
             for start in range(0, nline, step):
                 end = min(start + step, nline)
-                sol = sps.linalg.spsolve(Bbus[np.ix_(noslack, noref)].T,
-                                         Bf[np.ix_(luid[start:end], noref)].T).T
+                sol = sps.linalg.spsolve(A=Bbus[np.ix_(noslack, noref)].T,
+                                         b=Bf[np.ix_(luid[start:end], noref)].T,
+                                         permc_spec=permc_spec,
+                                         use_umfpack=use_umfpack).T
                 H[start:end, noslack] = sol
 
                 # show progress in percentage

{ltbams-0.9.8 → ltbams-0.9.10}/ams/interop/andes.py

@@ -11,6 +11,7 @@ from andes.utils.misc import elapsed
 from andes.system import System as andes_System
 
 from ams.io import input_formats
+from ams.shared import nan
 
 logger = logging.getLogger(__name__)
 
@@ -260,7 +261,7 @@ def parse_addfile(adsys, amsys, addfile):
                     except KeyError:  # set the most frequent string as the model name
                         split_list = []
                         for item in df[idxn].values:
-                            if item is None or np.nan:
+                            if item is None or nan:
                                 continue
                             try:
                                 split_list.append(item.split('_'))
@@ -295,7 +296,7 @@ def parse_addfile(adsys, amsys, addfile):
     # add dynamic models
     for name, df in df_models.items():
         # drop rows that all nan
-        df.replace(['', ' '], np.NaN, inplace=True)  # replace empty string with nan
+        df.replace(['', ' '], nan, inplace=True)  # replace empty string with nan
         df.dropna(axis=0, how='all', inplace=True)
         # if the dynamic model also exists in AMS, use AMS parameters for overlap
         if (name in amsys.models.keys()) and amsys.models[name].n > 0:
@@ -684,15 +685,15 @@ class Dynamic:
             cond_ads_stg_u = (mname_ads in ['StaticGen', 'PV', 'Sclak']) and (pname_ads == 'u')
             if cond_ads_stg_u and (self.is_tds):
                 # --- SynGen ---
-                u_sg = sa.SynGen.get(idx=link['syg_idx'].replace(np.NaN, None).to_list(),
+                u_sg = sa.SynGen.get(idx=link['syg_idx'].replace(nan, None).to_list(),
                                      src='u', attr='v',
                                      allow_none=True, default=0,)
                 # --- DG ---
-                u_dg = sa.DG.get(idx=link['dg_idx'].replace(np.NaN, None).to_list(),
+                u_dg = sa.DG.get(idx=link['dg_idx'].replace(nan, None).to_list(),
                                  src='u', attr='v',
                                  allow_none=True, default=0,)
                 # --- RenGen ---
-                u_rg = sa.RenGen.get(idx=link['rg_idx'].replace(np.NaN, None).to_list(),
+                u_rg = sa.RenGen.get(idx=link['rg_idx'].replace(nan, None).to_list(),
                                      src='u', attr='v',
                                      allow_none=True, default=0,)
                 # --- output ---
@@ -722,19 +723,19 @@ class Dynamic:
             cond_ads_stg_p = (mname_ads in ['StaticGen', 'PV', 'Sclak']) and (pname_ads == 'p')
             if cond_ads_stg_p and (self.is_tds):
                 # --- SynGen ---
-                p_sg = sa.SynGen.get(idx=link['syg_idx'].replace(np.NaN, None).to_list(),
+                p_sg = sa.SynGen.get(idx=link['syg_idx'].replace(nan, None).to_list(),
                                      src='Pe', attr='v',
                                      allow_none=True, default=0,)
                 # --- DG ---
-                Ie_dg = sa.DG.get(idx=link['dg_idx'].replace(np.NaN, None).to_list(),
+                Ie_dg = sa.DG.get(idx=link['dg_idx'].replace(nan, None).to_list(),
                                   src='Ipout_y', attr='v',
                                   allow_none=True, default=0,)
-                v_dg = sa.DG.get(idx=link['dg_idx'].replace(np.NaN, None).to_list(),
+                v_dg = sa.DG.get(idx=link['dg_idx'].replace(nan, None).to_list(),
                                  src='v', attr='v',
                                  allow_none=True, default=0,)
                 p_dg = Ie_dg * v_dg
                 # --- RenGen ---
-                p_rg = sa.RenGen.get(idx=link['rg_idx'].replace(np.NaN, None).to_list(),
+                p_rg = sa.RenGen.get(idx=link['rg_idx'].replace(nan, None).to_list(),
                                      src='Pe', attr='v',
                                      allow_none=True, default=0,)
                 # --- output ---

{ltbams-0.9.8 → ltbams-0.9.10}/ams/io/matpower.py

@@ -226,7 +226,7 @@ def mpc2system(mpc: dict, system) -> bool:
             system.add('GCost', gen=int(gen),
                        u=1, type=gctype,
                        idx=gcost_idx,
-                       name=f'GCost {gcost_idx}',
+                       name=None,
                        csu=startup, csd=shutdown,
                        c2=c2, c1=c1, c0=c0
                        )
@@ -235,7 +235,7 @@ def mpc2system(mpc: dict, system) -> bool:
     zone_id = np.unique(system.Bus.zone.v).astype(int)
     for zone in zone_id:
         zone_idx = f'ZONE_{zone}'
-        system.add('Region', idx=zone_idx, name=zone_idx)
+        system.add('Region', idx=zone_idx, name=None)
     bus_zone = system.Bus.zone.v
     bus_zone = [f'ZONE_{int(zone)}' for zone in bus_zone]
     system.Bus.zone.v = bus_zone
@@ -291,9 +291,11 @@ def system2mpc(system) -> dict:
                gen=np.zeros((system.PV.n + system.Slack.n, 21), dtype=np.float64),
                branch=np.zeros((system.Line.n, 17), dtype=np.float64),
                gencost=np.zeros((system.GCost.n, 7), dtype=np.float64),
-               bus_name=np.zeros((system.Bus.n, ), dtype=object),
                )
 
+    if system.Bus.name.v is not None:
+        mpc['bus_name'] = system.Bus.name.v
+
     base_mva = system.config.mva
 
     # --- bus ---
@@ -401,6 +403,4 @@ def system2mpc(system) -> dict:
     else:
         mpc.pop('gencost')
 
-    mpc['bus_name'] = np.array(system.Bus.name.v)
-
     return mpc

{ltbams-0.9.8 → ltbams-0.9.10}/ams/main.py

@@ -308,7 +308,7 @@ def run(filename, input_path='', verbose=20, mp_verbose=30,
         ncpu=NCPUS_PHYSICAL, pool=False,
         cli=False, shell=False, **kwargs):
     """
-    Entry point to run ANDES routines.
+    Entry point to run AMS routines.
 
     Parameters
     ----------
@@ -511,7 +511,7 @@ def demo(**kwargs):
 
 def versioninfo():
     """
-    Print version info for ANDES and dependencies.
+    Print version info for AMS and dependencies.
     """
 
     import numpy as np
@@ -563,7 +563,7 @@ def print_license():
 
 def edit_conf(edit_config: Optional[Union[str, bool]] = ''):
     """
-    Edit the Andes config file which occurs first in the search path.
+    Edit the AMS config file which occurs first in the search path.
 
     Parameters
     ----------
@@ -647,7 +647,7 @@ def save_conf(config_path=None, overwrite=None, **kwargs):
 # TODO: list AMS output files here
 def remove_output(recursive=False):
     """
-    Remove the outputs generated by Andes, including power flow reports
+    Remove the outputs generated by AMS, including power flow reports
     ``_out.txt``, time-domain list ``_out.lst`` and data ``_out.dat``,
     eigenvalue analysis report ``_eig.txt``.
 

{ltbams-0.9.8 → ltbams-0.9.10}/ams/opt/omodel.py

@@ -412,8 +412,10 @@ class Var(OptzBase):
 
     @v.setter
     def v(self, value):
-        # FIXME: is this safe?
-        self._v = value
+        if self.optz is None:
+            logger.info(f"Variable <{self.name}> is not initialized yet.")
+        else:
+            self.optz.value = value
 
     def get_idx(self):
         if self.is_group:
@@ -595,6 +597,10 @@ class Constraint(OptzBase):
         else:
             return self.optz._expr.value
 
+    @v.setter
+    def v(self, value):
+        raise AttributeError("Cannot set the value of the constraint.")
+
 
 class Objective(OptzBase):
     """
@@ -634,7 +640,6 @@ class Objective(OptzBase):
         OptzBase.__init__(self, name=name, info=info, unit=unit)
         self.e_str = e_str
         self.sense = sense
-        self._v = 0
         self.code = None
 
     @property
@@ -673,13 +678,12 @@ class Objective(OptzBase):
         """
         if self.optz is None:
             return None
-        out = self.om.obj.value
-        out = self._v if out is None else out
-        return out
+        else:
+            return self.optz.value
 
     @v.setter
     def v(self, value):
-        self._v = value
+        raise AttributeError("Cannot set the value of the objective function.")
 
     def parse(self, no_code=True):
         """

{ltbams-0.9.8 → ltbams-0.9.10}/ams/pypower/core/pips.py

@@ -1,22 +1,24 @@
 """
 Python Interior Point Solver (PIPS).
 """
-import logging  # NOQA
+import logging
 
-import numpy as np  # NOQA
-from numpy import flatnonzero as find  # NOQA
+import numpy as np
+from numpy import flatnonzero as find
 
-import scipy.sparse as sp  # NOQA
-from scipy.sparse import csr_matrix as c_sparse  # NOQA
+import scipy.sparse as sp
+from scipy.sparse import csr_matrix as c_sparse
 
-from andes.shared import deg2rad, rad2deg  # NOQA
+from andes.shared import deg2rad, rad2deg
 
-from ams.pypower.eps import EPS  # NOQA
-from ams.pypower.idx import IDX  # NOQA
-from ams.pypower.utils import sub2ind  # NOQA
-from ams.pypower.make import makeYbus  # NOQA
+from ams.shared import inf
+
+from ams.pypower.eps import EPS
+from ams.pypower.idx import IDX
+from ams.pypower.utils import sub2ind
+from ams.pypower.make import makeYbus
 from ams.pypower.routines.opffcns import (opf_costfcn, opf_consfcn,
-                                          opf_hessfcn,)  # NOQA
+                                          opf_hessfcn,)
 
 
 logger = logging.getLogger(__name__)
@@ -205,13 +207,13 @@ def pips(f_fcn, x0=None, A=None, l=None, u=None, xmin=None, xmax=None,
 
     # default argument values
     if l is None or len(l) == 0:
-        l = -np.Inf * np.ones(nA)
+        l = -inf * np.ones(nA)
     if u is None or len(u) == 0:
-        u = np.Inf * np.ones(nA)
+        u = inf * np.ones(nA)
     if xmin is None or len(xmin) == 0:
-        xmin = -np.Inf * np.ones(x0.shape[0])
+        xmin = -inf * np.ones(x0.shape[0])
     if xmax is None or len(xmax) == 0:
-        xmax = np.Inf * np.ones(x0.shape[0])
+        xmax = inf * np.ones(x0.shape[0])
     if gh_fcn is None:
         nonlinear = False
         gn = np.array([])
@@ -343,15 +345,15 @@ def pips(f_fcn, x0=None, A=None, l=None, u=None, xmin=None, xmax=None,
 
     maxh = np.zeros(1) if len(h) == 0 else max(h)
 
-    gnorm = np.linalg.norm(g, np.Inf) if len(g) else 0.0
-    lam_norm = np.linalg.norm(lam, np.Inf) if len(lam) else 0.0
-    mu_norm = np.linalg.norm(mu, np.Inf) if len(mu) else 0.0
-    znorm = np.linalg.norm(z, np.Inf) if len(z) else 0.0
+    gnorm = np.linalg.norm(g, inf) if len(g) else 0.0
+    lam_norm = np.linalg.norm(lam, inf) if len(lam) else 0.0
+    mu_norm = np.linalg.norm(mu, inf) if len(mu) else 0.0
+    znorm = np.linalg.norm(z, inf) if len(z) else 0.0
     feascond = \
-        max([gnorm, maxh]) / (1 + max([np.linalg.norm(x, np.Inf), znorm]))
+        max([gnorm, maxh]) / (1 + max([np.linalg.norm(x, inf), znorm]))
     gradcond = \
-        np.linalg.norm(Lx, np.Inf) / (1 + max([lam_norm, mu_norm]))
-    compcond = np.dot(z, mu) / (1 + np.linalg.norm(x, np.Inf))
+        np.linalg.norm(Lx, inf) / (1 + max([lam_norm, mu_norm]))
+    compcond = np.dot(z, mu) / (1 + np.linalg.norm(x, inf))
     costcond = np.abs(f - f0) / (1 + np.abs(f0))
 
     # save history
@@ -467,14 +469,14 @@ def pips(f_fcn, x0=None, A=None, l=None, u=None, xmin=None, xmax=None,
 
             maxh1 = np.zeros(1) if len(h1) == 0 else max(h1)
 
-            g1norm = np.linalg.norm(g1, np.Inf) if len(g1) else 0.0
-            lam1_norm = np.linalg.norm(lam, np.Inf) if len(lam) else 0.0
-            mu1_norm = np.linalg.norm(mu, np.Inf) if len(mu) else 0.0
-            z1norm = np.linalg.norm(z, np.Inf) if len(z) else 0.0
+            g1norm = np.linalg.norm(g1, inf) if len(g1) else 0.0
+            lam1_norm = np.linalg.norm(lam, inf) if len(lam) else 0.0
+            mu1_norm = np.linalg.norm(mu, inf) if len(mu) else 0.0
+            z1norm = np.linalg.norm(z, inf) if len(z) else 0.0
 
             feascond1 = max([g1norm, maxh1]) / \
-                (1 + max([np.linalg.norm(x1, np.Inf), z1norm]))
-            gradcond1 = np.linalg.norm(Lx1, np.Inf) / (1 + max([lam1_norm, mu1_norm]))
+                (1 + max([np.linalg.norm(x1, inf), z1norm]))
+            gradcond1 = np.linalg.norm(Lx1, inf) / (1 + max([lam1_norm, mu1_norm]))
 
             if (feascond1 > feascond) and (gradcond1 > gradcond):
                 sc = True
@@ -560,15 +562,15 @@ def pips(f_fcn, x0=None, A=None, l=None, u=None, xmin=None, xmax=None,
         else:
             maxh = max(h)
 
-        gnorm = np.linalg.norm(g, np.Inf) if len(g) else 0.0
-        lam_norm = np.linalg.norm(lam, np.Inf) if len(lam) else 0.0
-        mu_norm = np.linalg.norm(mu, np.Inf) if len(mu) else 0.0
-        znorm = np.linalg.norm(z, np.Inf) if len(z) else 0.0
+        gnorm = np.linalg.norm(g, inf) if len(g) else 0.0
+        lam_norm = np.linalg.norm(lam, inf) if len(lam) else 0.0
+        mu_norm = np.linalg.norm(mu, inf) if len(mu) else 0.0
+        znorm = np.linalg.norm(z, inf) if len(z) else 0.0
         feascond = \
-            max([gnorm, maxh]) / (1 + max([np.linalg.norm(x, np.Inf), znorm]))
+            max([gnorm, maxh]) / (1 + max([np.linalg.norm(x, inf), znorm]))
         gradcond = \
-            np.linalg.norm(Lx, np.Inf) / (1 + max([lam_norm, mu_norm]))
-        compcond = np.dot(z, mu) / (1 + np.linalg.norm(x, np.Inf))
+            np.linalg.norm(Lx, inf) / (1 + max([lam_norm, mu_norm]))
+        compcond = np.dot(z, mu) / (1 + np.linalg.norm(x, inf))
         costcond = float(np.abs(f - f0) / (1 + np.abs(f0)))
 
         hist.append({'feascond': feascond, 'gradcond': gradcond,
@@ -775,8 +777,8 @@ def pipsopf_solver(om, ppopt, out_opt=None):
 
     # try to select an interior initial point
     ll, uu = xmin.copy(), xmax.copy()
-    ll[xmin == -np.Inf] = -1e10  # replace Inf with numerical proxies
-    uu[xmax == np.Inf] = 1e10
+    ll[xmin == -inf] = -1e10  # replace Inf with numerical proxies
+    uu[xmax == inf] = 1e10
     x0 = (ll + uu) / 2
     Varefs = bus[bus[:, IDX.bus.BUS_TYPE] == IDX.bus.REF, IDX.bus.VA] * deg2rad
     # angles set to first reference angle

{ltbams-0.9.8 → ltbams-0.9.10}/ams/pypower/core/ppoption.py

@@ -7,6 +7,8 @@
 
 import numpy as np
 
+from ams.shared import inf
+
 
 PF_OPTIONS = [
     ('pf_alg', 1, '''power flow algorithm:
@@ -184,7 +186,7 @@ GUROBI_OPTIONS = [
 3 - concurrent (LP only)
 4 - deterministic concurrent (LP only)
 '''),
-    ('grb_timelimit', np.Inf, 'maximum time allowed for solver (TimeLimit)'),
+    ('grb_timelimit', inf, 'maximum time allowed for solver (TimeLimit)'),
     ('grb_threads', 0, '(auto) maximum number of threads to use (Threads)'),
     ('grb_opt', 0, 'See gurobi_options() for details')
 ]