fixed loki + some opti changes

backend/src/optimization/optimizer.py

@@ -87,7 +87,7 @@ class Optimizer:
 
         # inequality matrix and vector
         A_ub = np.zeros(L*L, dtype=np.int16)
-        b_ub = round(max_time*self.overshoot)
+        b_ub = round(max_time*(1.1+max_time*self.overshoot))
 
         for i, spot1 in enumerate(landmarks) :
             c[i] = spot1.attractiveness
@@ -489,6 +489,17 @@ class Optimizer:
         return L
 
 
+    def warm_start(self, x: list[pl.LpVariable], L: int) :
+
+        for i in range(L*L) :
+            x[i].setInitialValue(0)
+            
+        x[1].setInitialValue(1)
+        x[2*L-1].setInitialValue(1)
+
+        return x
+
+
     def pre_processing(self, L: int, landmarks: list[Landmark], max_time: int, max_landmarks: int | None) :
         """
         Preprocesses the optimization problem by setting up constraints and variables for the tour optimization.
@@ -539,6 +550,7 @@ class Optimizer:
         self.respect_order(prob, x, L)                   # Respect order of visit (only works when max_time is limiting factor)
         self.respect_user_must(prob, x, L, landmarks)    # Force to do/avoid landmarks set by user.
 
+        # return prob, self.warm_start(x, L)
         return prob, x
 
 
@@ -561,7 +573,7 @@ class Optimizer:
         prob, x = self.pre_processing(L, landmarks, max_time, max_landmarks)
 
         # Solve the problem and extract results.
-        prob.solve(pl.PULP_CBC_CMD(msg=False, gapRel=0.1))
+        prob.solve(pl.PULP_CBC_CMD(msg=False, gapRel=0.1, timeLimit=10, warmStart=False))
         status = pl.LpStatus[prob.status]
         solution = [pl.value(var) for var in x]  # The values of the decision variables (will be 0 or 1)