started with TDA

schedTests/LiuAndLaylandBound.py

@@ -19,9 +19,9 @@ def test(tasks):
     U_lub = n * ((2 ** (1 / n)) - 1)
     
     # for fewer tasks than 10, we use the exact computed least upper bound
-    if n < 10:
-        return U <= U_lub
+   #  if n < 10:
+   #     return U <= U_lub
 
     # from 10 tasks up unlimited, we use the limes of n(2 ** 1/n - 1)
-    return U <= np.log(2) # round to 0.7 ?
+    return U <= U_lub # np.log(2) 
 

schedTests/TimeDemandAnalysis.py

@@ -13,15 +13,42 @@ import include.TasksHelper as TH
 # C_i is accessed as: tasks[i][2]
 # The number of tasks can be accessed as: tasks.shape[0]
 
-#The Time Demand Analysis Test
+# The Time Demand Analysis Test
+
+
 def test(tasks):
-    #Sorting Taskset by Period/Deadline
-    #This makes implementing TDA a lot easier
+    # Sorting Taskset by Period/Deadline
+    # This makes implementing TDA a lot easier
     shape = tasks.shape
     sortedtasks = tasks[tasks[:, 0].argsort()]
 
-    #####################
-    #YOUR CODE GOES HERE#
-    #####################
+    print("\n======= TASK SET =======")
+    # For each tasks in the ordered set
+    for i in range(len(sortedtasks)):
 
-    return False
+        print(f'Task #{i} {tasks[i]}:')
+
+        # calculate the time points for the demand function
+        # t = j * P_k for k = 1, 2,...i and j = 1, 2,...,math.ceil(P_i / P_k)
+        list_of_t = [TH.P_i(sortedtasks, k) * j for k in range(1, i)
+                     for j in range(1, int(np.ceil(TH.D_i(sortedtasks, i) / TH.P_i(sortedtasks, k))))]
+
+        # print(f'\t list of ts: {list_of_t}')
+
+        # at any time t between 0 and and TH.P_i
+        for t in np.sort(list_of_t):
+        # for t in range(TH.C_i(sortedtasks, i), TH.P_i(sortedtasks, i)+1, TH.P_i(sortedtasks, i)):
+
+            print(f'\tTime-Demand for t ({t}): {time_demand_func(sortedtasks, i, t)} ')
+
+            # if the demand for CPU time of task i exceeds the available time t
+            if time_demand_func(sortedtasks, i, t) > t:
+                return False # then the task i will not meet its deadline, hence taskset not schedulable
+    return True
+
+
+def time_demand_func(tasks, i, delta=0):
+    sum = 0
+    for k in range(1, i-1):
+        sum += math.ceil(delta / TH.P_i(tasks, k)) * TH.C_i(tasks, k)
+    return TH.C_i(tasks, i) + sum