Upload New File

schedule.py

+# coding=utf-8
+
+import os
+import pandas as pd
+from pulp import LpMinimize, LpProblem, lpSum, LpVariable
+import matplotlib.pyplot as plt
+import numpy as np
+
+cur_dir = os.path.dirname(os.path.abspath("__file__"))
+charts_dir = os.path.join(cur_dir, "charts")
+
+
+def read_input_data() -> (list, list, list, list):
+    """Read user task information from Excel file
+    """
+    excel_file = pd.read_excel(
+        os.path.join(cur_dir, "COMP3217CW2Input.xlsx"),
+        sheet_name="User & Task ID"
+    )
+    task_name = excel_file['User & Task ID'].tolist()
+    ready_time = excel_file['Ready Time'].tolist()
+    deadline = excel_file['Deadline'].tolist()
+    max_energy_per_hour = excel_file['Maximum scheduled energy per hour'].tolist()
+    energy_demand = excel_file['Energy Demand'].tolist()
+    all_tasks = list()
+    all_task_names = []
+
+    for k in range(len(ready_time)):
+        task = list()
+        task.append(ready_time[k])
+        task.append(deadline[k])
+        task.append(max_energy_per_hour[k])
+        task.append(energy_demand[k])
+        all_task_names.append(task_name[k])
+
+        all_tasks.append(task)
+
+    # Reading Testing Data Output
+    test_data = pd.read_csv('TestingResults.txt', header=None)
+    y_labels = test_data[24].tolist()
+    test_data = test_data.drop(24, axis=1)
+    x_data = test_data.values.tolist()
+
+    return all_tasks, all_task_names, x_data, y_labels
+
+
+def create_LP_model(tasks: list, task_names: list) -> LpProblem:
+    """Create an LP model for the problem
+    """
+
+    # some variables
+    task_vars = []
+    c = []
+    eq = []
+
+    # create LP problem model for Minimization
+    model = LpProblem(name="scheduling-problem", sense=LpMinimize)
+
+    # Loop through list of tasks
+    for i, task in enumerate(tasks):
+        temp_list = []
+        # Loop between ready_time and deadline for each task
+        # Creates LP variables with given constraints and unique names
+        for j in range(task[0], task[1] + 1):
+            x = LpVariable(name=task_names[i] + '_' + str(j), lowBound=0, upBound=task[2])
+            temp_list.append(x)
+        task_vars.append(temp_list)
+
+    # Create objective function for price (to minimize) and add to the model
+    for i, task in enumerate(tasks):
+        for var in task_vars[i]:
+            price = price_list[int(var.name.split('_')[2])]
+            c.append(price * var)
+    model += lpSum(c)
+
+    # Add additional constraints to the model
+    for i, task in enumerate(tasks):
+        temp_list = []
+        for var in task_vars[i]:
+            temp_list.append(var)
+        eq.append(temp_list)
+        model += lpSum(temp_list) == task[3]
+
+    # Return model to be solved in main function
+    return model
+
+
+def plot(model: LpProblem, count: int) -> list:
+    """Draw the charts: hourly energy usage for users
+    """
+    hours = [str(x) for x in range(0, 24)]
+    pos = np.arange(len(hours))
+    users = ["user1", "user2", "user3", "user4", "user5"]
+    plot_list = []
+
+    # Create lists to plot usage
+    for user in users:
+        temp_list = []
+        for hour in hours:
+            hour_list_temp = []
+            task_count = 0
+            for var in model.variables():
+                if user == var.name.split('_')[0] and str(hour) == var.name.split('_')[2]:
+                    task_count += 1
+                    hour_list_temp.append(var.value())
+            temp_list.append(sum(hour_list_temp))
+        plot_list.append(temp_list)
+
+    # Show as bar stack chart
+    plt.bar(pos, plot_list[0], edgecolor='black', bottom=0)
+    plt.bar(pos, plot_list[1], edgecolor='black', bottom=np.array(plot_list[0]))
+    plt.bar(pos, plot_list[2], edgecolor='black',
+            bottom=np.array(plot_list[0]) + np.array(plot_list[1]))
+    plt.bar(pos, plot_list[3], edgecolor='black',
+            bottom=np.array(plot_list[0]) + np.array(plot_list[1]) + np.array(plot_list[2]))
+    plt.bar(pos, plot_list[4], edgecolor='black',
+            bottom=np.array(plot_list[0]) + np.array(plot_list[1]) + np.array(plot_list[2]) + np.array(plot_list[3]))
+
+    # Add sum value line chart
+    sum_value_list = list()
+    for i, _ in enumerate(plot_list[0]):
+        sum_value_list.append(plot_list[0][i] + plot_list[1][i] + plot_list[2][i] + plot_list[3][i] + plot_list[4][i])
+
+    plt.xticks(pos, hours)
+    plt.xlabel("Time [Hour]")
+    plt.ylabel("Energy Usage [kW]")
+    plt.title("Energy Usage Per Hour For All Five Users\nLine {}".format(count))
+    plt.legend(users, loc=0)
+    # plt.show()
+    plt.savefig("{}/bar_stack_{}.png".format(charts_dir, count))
+    plt.clf()
+
+    pos_str = [str(p) for p in pos]
+    plt.plot(pos_str, sum_value_list)
+    plt.xlabel("Time [Hour]")
+    plt.ylabel("Energy Usage [kW]")
+    plt.title("Energy Usage Per Hour For All Five Users\nLine {}".format(count))
+    plt.savefig("{}/line_{}.png".format(charts_dir, count))
+    plt.clf()
+
+    return plot_list
+
+
+if __name__ == "__main__":
+    tasks, task_names, x_data, y_labels = read_input_data()
+
+    for ind, price_list in enumerate(x_data):
+        # Schedule and plot abnormal guideline pricing curves
+        if y_labels[ind] == 1:
+            # Solve returned LP model for scheduling solution
+            model = create_LP_model(tasks, task_names)
+            answer = model.solve()
+            # draw hourly usage for scheduling solution
+            plot(model, ind + 1)