rename sigma to slack

README.md

@@ -8,7 +8,7 @@ _License_: MIT
 
 _Other contributors_:
     - Many thanks to [Edgar W.](https://github.com/techniccontroller) for spotting out a bug that
-      let the slack variables  `sigma_u` and `sigma_y` assume any value when the value of the 
+      let the slack variables  `slack_u` and `slack_y` assume any value when the value of the 
       corresponding regularizer was `0`.
 
 ![Closed loop results](examples/images/example_siso_pulley.png "Pulley system")

examples/example_mimo_twotank.py

@@ -80,8 +80,8 @@ def constraints_callback(u: cp.Variable, y: cp.Variable) -> List[Constraint]:
 
     # Plot curve
     data = sys.get_all_samples()
-    ax[0].plot(data.y, label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
-    ax[1].plot(data.u, label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
+    ax[0].plot(data.y[T_INI:], label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
+    ax[1].plot(data.u[T_INI:], label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
 
 ax[0].set_ylim(0, 1.5)
 ax[1].set_ylim(-1.2, 1.2)

examples/example_siso_pulley.py

@@ -82,8 +82,8 @@ def constraints_callback(u: cp.Variable, y: cp.Variable) -> List[Constraint]:
 
     # Plot curve
     data = sys.get_all_samples()
-    ax[0].plot(data.y, label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
-    ax[1].plot(data.u, label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
+    ax[0].plot(data.y[T_INI:], label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
+    ax[1].plot(data.u[T_INI:], label=f'$s={s}, T={T}, T_i={T_INI}, N={HORIZON}$')
 
 ax[0].set_ylim(0, 2)
 ax[1].set_ylim(-4, 4)

pydeepc/__init__.py

@@ -2,7 +2,7 @@
 from .utils import *
 
 __author__ = 'Alessio Russo - [email protected]'
-__version__ = '0.3.4'
+__version__ = '0.3.5'
 __url__ = 'https://github.com/rssalessio/PyDeePC'
 __info__ = {
     'version': __version__,

pydeepc/deepc.py

@@ -111,8 +111,8 @@ def build_problem(self,
         u = cp.Variable(shape=(self.M * self.horizon), name='u')
         y = cp.Variable(shape=(self.P * self.horizon), name='y')
         g = cp.Variable(shape=(self.T - self.Tini - self.horizon + 1), name='g')
-        sigma_y = cp.Variable(shape=(self.Tini * self.P), name='sigma_y')
-        sigma_u = cp.Variable(shape=(self.Tini * self.M), name='sigma_u')
+        slack_y = cp.Variable(shape=(self.Tini * self.P), name='slack_y')
+        slack_u = cp.Variable(shape=(self.Tini * self.M), name='slack_u')
 
         Up, Yp, Uf, Yf = self.Up, self.Yp, self.Uf, self.Yf
 
@@ -125,15 +125,15 @@ def build_problem(self,
             I_min_P = I - (ZpInv@ Zp)
 
         A = np.vstack([Up, Yp, Uf, Yf])
-        b = cp.hstack([uini + sigma_u, yini + sigma_y, u, y])
+        b = cp.hstack([uini + slack_u, yini + slack_y, u, y])
 
         # Build constraints
         constraints = [A @ g == b]
 
         if math.isclose(lambda_y, 0):
-            constraints.append(cp.norm(sigma_y, 2) <= DeePC._SMALL_NUMBER)
+            constraints.append(cp.norm(slack_y, 2) <= DeePC._SMALL_NUMBER)
         if math.isclose(lambda_u, 0):
-            constraints.append(cp.norm(sigma_u, 2) <= DeePC._SMALL_NUMBER)
+            constraints.append(cp.norm(slack_u, 2) <= DeePC._SMALL_NUMBER)
 
         # u, y = self.Uf @ g, self.Yf @ g
         u = cp.reshape(u, (self.horizon, self.M))
@@ -155,9 +155,9 @@ def build_problem(self,
 
         # Add regularizers
         _regularizers = lambda_g * cp.norm(g, p=1) if lambda_g > DeePC._SMALL_NUMBER else 0
-        _regularizers += lambda_y * cp.norm(sigma_y, p=1) if lambda_y > DeePC._SMALL_NUMBER else 0
+        _regularizers += lambda_y * cp.norm(slack_y, p=1) if lambda_y > DeePC._SMALL_NUMBER else 0
         _regularizers += lambda_proj * cp.norm(I_min_P @ g) if lambda_proj > DeePC._SMALL_NUMBER  else 0
-        _regularizers += lambda_u * cp.norm(sigma_u, p=1) if lambda_u > DeePC._SMALL_NUMBER else 0
+        _regularizers += lambda_u * cp.norm(slack_u, p=1) if lambda_u > DeePC._SMALL_NUMBER else 0
 
         problem_loss = _loss + _regularizers
 
@@ -170,7 +170,8 @@ def build_problem(self,
             raise Exception(f'Error while constructing the DeePC problem. Details: {e}')
 
         self.optimization_problem = OptimizationProblem(
-            variables = OptimizationProblemVariables(u_ini = uini, y_ini = yini, u = u, y = y, g = g, sigma_y = sigma_y),
+            variables = OptimizationProblemVariables(
+                u_ini = uini, y_ini = yini, u = u, y = y, g = g, slack_y = slack_y, slack_u = slack_u),
             constraints = constraints,
             objective_function = problem_loss,
             problem = problem

pydeepc/utils.py

@@ -13,8 +13,8 @@ class OptimizationProblemVariables(NamedTuple):
     u: Union[Variable, Parameter]
     y: Union[Variable, Parameter]
     g: Union[Variable, Parameter]
-    sigma_y: Union[Variable, Parameter]
-    sigma_u: Union[Variable, Parameter]
+    slack_y: Union[Variable, Parameter]
+    slack_u: Union[Variable, Parameter]
 
 
 class OptimizationProblem(NamedTuple):

setup.py

@@ -4,7 +4,7 @@
 
 setup(name = 'PyDeePC',
     packages=find_packages(),
-    version = '0.3.4',
+    version = '0.3.5',
     description = 'Python library for Data-Enabled Predictive Control',
     url = 'https://github.com/rssalessio/PyDeePC',
     author = 'Alessio Russo',