def main():
##############################
# DATA LOADING
##############################
with open('classes.csv', mode="r") as class_csv:
# read in the courses
class_reader = csv.DictReader(class_csv, delimiter=",")
courses = []
for row in class_reader:
c = Course(row["Course"], [
row["Monday"] == "x", row["Tuesday"] == "x", row["Wednesday"]
== "x", row["Thursday"] == "x", row["Friday"] == "x"
], row["Sections"], row["Hours"], row["Preferred Time"],
row["Preferred Room Type"], row["Students Per Section"])
courses.append(c)
print("Course list loaded")
with open('rooms.csv', mode="r") as room_csv:
# read in the rooms
room_reader = csv.DictReader(room_csv, delimiter=",")
rooms = []
for row in room_reader:
r = Room(row["Room"], row["Capacity"], row["Type"])
rooms.append(r)
print("Room list loaded")
# print(rooms[0])
##############################
# SOLUTION GENERATION
##############################
def generate_solution():
# Create 2880 room resources
available = AvailabeRooms()
# loop through all the days
for day in DAYS:
# add availability for every room
for room in rooms:
# 8:00 (0), 8:30 (1), 9:00 (2), 9:30 (3),
# 10:00 (4), 10:30 (5), 11:00 (6), 11:30 (7),
# 12:00 (8), 12:30 (9), 1:00 (10), 1:30 (11),
# 2:00 (12), 2:30 (13), 3:00 (14), 3:30 (15),
# 4:00 (16), 4:30 (17)
for i in range(0, 18):
available.add(RoomResource(room, day, i))
# randomize course order and create each section
# print("Randomize courses")
shuffle(courses)
sections = []
for course in courses:
for i in range(course.sections):
sections.append(course)
# print("Add to Solution")
cas = []
outside_building = []
for course in sections:
# get enough room resources required for the course
room_resources = available.pull_out_resources(course)
if len(room_resources) == 0:
outside_building.append(course)
else:
cas.append(CourseAssignment(room_resources, course))
return Solution(cas, outside_building, available)
##############################
# MODEL TESTING
##############################
run_number = 0
for options in RUN_OPTIONS:
run_number += 1
print(f"------------RUN {run_number}--------------")
print(
f"-- Population: {options.num_solutions} - Elite: {options.pct_elite}"
)
print(
f"-- Crossover: {options.pct_cross} - Mutation: {options.pct_mut}")
print("-------------------------------")
solutions = []
# generate a list of solutions to provide to the generation
for i in range(options.num_solutions):
solutions.append(generate_solution())
# create generation with solutions and percentages from run options
base_gen = Generation(solutions, options)
# order solutions in the generation based on fitness
base_gen.sort_solutions_fitness()
##############################
# GENERATION IMPROVEMENT
##############################
gen_avg = []
gen_number = 1
moving_avg = 999
# loop until moving average stabilizes below 1.0 for last 10 items
print(f"Generation, Mean Fitness, Max Fitness, Min Fitness")
while moving_avg > 1:
# create a new generation
# keep elites
new_gen = Generation(base_gen.get_elites(), options)
gen_avg.append(new_gen.avg_fitness())
print(
f"{gen_number}, {new_gen.avg_fitness()}, {new_gen.max_fitness()}, {new_gen.min_fitness()}"
)
# use crossover percentage from the solution set to fill in the rest of the new generation
new_gen.crossover_solutions(base_gen)
# mutate the % in the new generation
new_gen.mutate_solutions()
# calculate moving average
avgs = gen_avg[-10:]
avg_diffs = []
if len(avgs) >= 10:
for i in range(len(avgs) - 1):
avg_diffs.append(abs(avgs[i] - avgs[i + 1]))
moving_avg = mean(avg_diffs)
# reset the base_gen to create another generation on next loop
base_gen = new_gen
gen_number += 1
##############################
# OUTPUT RESULTS
##############################
new_gen.sort_solutions_fitness()
best_solution = new_gen.solutions[0]
print()
print("Best solution:")
print("Course, Days, Room, Start Time, End Time")
for ca in best_solution.course_assignments:
print(
f"{ca.course.name}, {ca.day}, {ca.room.number}, {ca.convert_to_time_range()}"
)
