def calc_score(courses, rooms, course_names, matrix):
"""
Score calculation.
"""
# iterate through courses and set all scores to zero
for course in courses:
course.goodbad = 0
# calculate points for all categories and add them
points = 0
points += sc.matrix_checker(courses, course_names, matrix) + sc.order_checker(courses)
points += sc.student_checker(rooms, courses, course_names)
dist_bonus, dist_malus = sc.distribution_checker(courses)
points += dist_bonus + dist_malus
points += sc.evening_checker(rooms, courses, course_names)
return points
def initial_population():
"""
Initial population.
"""
# store different schedules
scheduled_rooms = [[] for i in range(5)]
scheduled_courses = [[] for i in range(5)]
scheduled_students = [[] for i in range(5)]
scores = [0 for i in range(5)]
matrix = []
course_names = []
matrixfile = open("data/matrix.csv", 'r')
coursenamefile = open("data/vakken.txt", 'r')
# read matrix
for line in matrixfile:
matrix.append(line.split(";"))
for line in coursenamefile:
course_names.append(line.split(";")[0])
# schedule 5 times
for i in range(5):
#read files
coursefile = open("data/vakken.txt", 'r')
roomfile = open("data/lokalen.txt", 'r')
studentfile = open("data/studentenenvakken.csv", 'r', errors='ignore')
for line in coursefile:
info = line.split(";")
scheduled_courses[i].append(inf.Course(info[0], int(info[1]), int(info[2]), int(info[3]),\
int(info[4]), int(info[5]), int(info[7])))
for line in roomfile:
info = line.split(",")
scheduled_rooms[i].append(inf.Room_info(info[0], int(info[1])))
# create evening timeslot in largest room
big_room_cap = 0
for room in scheduled_rooms[i]:
if room.cap > big_room_cap:
big_room_cap = room.cap
big_room = room
for day in big_room.days:
day.hours.append(inf.Hour())
for line in studentfile:
student_info = line.strip("\n").split(";")
student_courses = []
for coursename in student_info[3:]:
if len(coursename) > 2:
student_courses.append(coursename)
scheduled_students[i].append(
inf.Student(student_info[0], student_info[1], student_info[2],
student_courses))
# remove headers in students
scheduled_students[i] = scheduled_students[i][1:]
# schedule
day_sch.total_schedule(scheduled_rooms[i], scheduled_courses[i],
course_names, matrix)
# calculate score and apply hillclimber
scores[i] = sc.matrix_checker(scheduled_courses[i], course_names,
matrix) + sc.order_checker(
scheduled_courses[i])
scores[i] += sc.student_checker(scheduled_rooms[i],
scheduled_courses[i], course_names)
bonus, malus = sc.distribution_checker(scheduled_courses[i])
scores[i] += bonus + malus
scores[i] += sc.evening_checker(scheduled_rooms[i],
scheduled_courses[i], course_names)
scores[i] = hill.random_climber(scheduled_courses[i],
scheduled_rooms[i], course_names, 1000,
scores[i], matrix)
# schedule students
stu.distribute_all_students(scheduled_students[i], scheduled_rooms[i],
scheduled_courses[i], course_names)
# Save score for hillclimber
student_bonus, student_malus = sc.student_score(scheduled_students[i])
student_score = student_bonus + student_malus
# pre filter the relevant courses
student_courses = []
for course in scheduled_courses[i]:
poss_group_ids = []
for activity in course.activities:
if activity.group_id not in poss_group_ids and activity.group_id != "x":
poss_group_ids.append(activity.group_id)
if len(poss_group_ids) > 1:
student_courses.append([course, poss_group_ids])
# student hillclimber
scores[i] += sthl.students_hillclimber(student_courses,
scheduled_students[i],
student_score, 100)
return scheduled_rooms, scheduled_courses, scheduled_students, scores, course_names, matrix
def mutate_schedule(rooms, courses, students, score, matrix, course_names):
"""
Mutate a fit schedule to 10 different schedules. Keep the orignal schedule.
Furthermore, for 9 times: unschedule a random subset of the courses and
reschedule this subset.
"""
# store mutated schedules
new_rooms = [[inf.Room_info(room.name, room.cap) for room in rooms] \
for j in range(10)]
new_courses = [[inf.Course(course.name, course.hoorcolleges, \
course.werkcolleges, course.max_werkcolleges, course.practica, \
course.max_practica, course.e_students) for course in courses] \
for j in range(10)]
new_students = [[inf.Student(student.surname, student.name, \
student.student_number, student.courses) for student in students] \
for j in range(10)]
new_scores = [0 for j in range(10)]
# create evening timeslot in largest room
for n_rooms in new_rooms:
big_room_cap = 0
for room in n_rooms:
if room.cap > big_room_cap:
big_room_cap = room.cap
big_room = room
for day in big_room.days:
day.hours.append(inf.Hour())
# copy orignal schedule onto mutations
for k in range(10):
new_rooms[k], new_courses[k], new_students[k] = bas_sch.copy_schedule(rooms, \
courses, students, new_rooms[k], new_courses[k], new_students[k])
# keep original schedule
new_scores[0] = score
# mutate 9 times
for i in range(1, 10):
# select random subset of scheduled courses
indices = rd.sample(range(len(new_courses[i])), rd.randint(0, \
len(new_courses[i]) - 7))
# store names of subset
unsched_names = []
# unschedule subset
for j in indices:
unsched_names.append(new_courses[i][j].name)
clear_schedule_of_course(new_rooms[i], new_courses[i][j], \
new_students[i])
# schedule all required classes
schedulings = 0
for j in indices:
schedulings += day_sch.course_scheduler(new_courses[i][j], new_rooms[i],\
new_courses[i], course_names, matrix)
print("Schedulings:", schedulings)
# schedule students' unscheduled groups
for student in new_students[i]:
for coursename in student.courses:
if coursename in unsched_names:
course = new_courses[i][course_names.index(coursename)]
poss_group_ids = []
student_id = ""
# get possible groups of course for student and choose one randomly
for activity in course.activities:
if activity.group_id != "x" and len(
activity.students) < activity.capacity:
poss_group_ids.append(activity.group_id)
if len(poss_group_ids) > 0:
student_id = rd.choice(poss_group_ids)
student.group_id[student.courses.index(
course.name)] = student_id
# schedule student
for activity in course.activities:
if activity.id == "Hoorcollege" or activity.group_id == student_id:
activity.students.append(student.student_number)
student.dates[student.courses.index(
coursename)].append(activity.date)
# calculate score
new_scores[i] = sc.matrix_checker(new_courses[i], course_names, \
matrix) + sc.order_checker(new_courses[i])
new_scores[i] += sc.student_checker(new_rooms[i], new_courses[i],
course_names)
bonus, malus = sc.distribution_checker(new_courses[i])
new_scores[i] += bonus + malus + sc.evening_checker(new_rooms[i], \
new_courses[i], course_names)
student_bonus, student_malus = sc.student_score(new_students[i])
new_scores[i] += student_bonus + student_malus
return new_rooms, new_courses, new_students, new_scores
for room in rooms:
if room.cap > big_room_cap:
big_room_cap = room.cap
big_room = room
for day in big_room.days:
day.hours.append(inf.Hour())
course_names = [course.name for course in courses]
# return total score of schedule using the score calculator
day_sch.total_schedule(rooms, courses, course_names, matrix)
score = sc.matrix_checker(courses, course_names,
matrix) + sc.order_checker(courses)
score += sc.student_checker(rooms, courses, course_names)
bonus, malus = sc.distribution_checker(courses)
score += bonus + malus
score += sc.evening_checker(rooms, courses, course_names)
# print scores before hillclimber
if not course_optim_type == "none":
print("Courses score before optimization:", score)
# for course in courses:
# if course.goodbad < - 1000:
# score = hill.course_climber(courses[0], courses, rooms, course_names, 1000, score, matrix)
# print("Score after 1 course_climb: ", score)
# run hillclimber if specified by user
if course_optim_type == "hillclimber":
def search(rooms, courses, course_names, students, matrix):
"""
Searches for painpoints in the case.
"""
student_numbers = [student.student_number for student in students]
student_weakness = [0 for number in student_numbers]
course_weakness = [0 for coursename in course_names]
for i in range(100):
# schedule
day_sch.total_schedule(rooms, courses, course_names, matrix)
# calculate score and apply hillclimber
score = sc.matrix_checker(courses, course_names,
matrix) + sc.order_checker(courses)
score += sc.student_checker(rooms, courses, course_names)
bonus, malus = sc.distribution_checker(courses)
score += bonus + malus
score += sc.evening_checker(rooms, courses, course_names)
score = hill.random_climber(courses, rooms, course_names, 1000, score,
matrix)
# schedule students
stu.distribute_all_students(students, rooms, courses, course_names)
# save score for hillclimber
student_bonus, student_malus = sc.student_score(students)
student_score = student_bonus + student_malus
# pre filter the relevant courses
student_courses = []
for course in courses:
poss_group_ids = []
for activity in course.activities:
if activity.group_id not in poss_group_ids and activity.group_id != "x":
poss_group_ids.append(activity.group_id)
if len(poss_group_ids) > 1:
student_courses.append([course, poss_group_ids])
# student hillclimber
student_climb_score = sthl.students_hillclimber(
student_courses, students, student_score, 100)
# update statistics
for course in courses:
if course.goodbad < 0:
course_weakness[course_names.index(course.name)] += 1
for student in students:
if student.goodbad < -12:
student_weakness[student_numbers.index(
student.student_number)] += 1
bas_sch.print_schedule(rooms)
bas_sch.clear_schedule(rooms, courses)
bas_sch.clear_students(students)
cprint(i, "blue")
# print painpoints
for f in range(len(student_numbers)):
print(student_numbers[f] + ":", student_weakness[f])
for j in range(len(course_names)):
print(course_names[j] + ":", course_weakness[j])