How to implement a timetable scheduling using GA with toolbox?

Hello everyone.

First of all, i would like to express my gratitude to all that helped me.

Im trying to implement a GA for timetable scheduling for my university using the toolbox in matlab. However, i have no idea how and where to start. Please leave some suggestions on where can i get help in. Any kind of material or journals suggestions are welcome as well. Thank you.

[Walter Roberson]

First you have to identify what it is you plan to minimize.

[yiwen yim]

May i ask for further explanation? Thank you.

[Walter Roberson]

Suppose I propose two different timetables, both of which meet all constraints about conflicts. How do you measure which of the two timetables is "better" ?

GA is not a tool for searching for a some configuration that meets the constraints and then just stopping. GA is a minimizer, a tool for looking for the best configuration according to some numeric measure.

For example, one of the proposed timetables might suggest 12 classes on the same day, which might in some measure be considered good for that one day because it cuts down on the amount of travel to and from university, but you would probably also want to penalize it for being too tiring. Is a schedule that goes "1 hour class" "1 hour break" "1 hour class" "done for the day" less tiring than a schedule that goes "1 hour class" "1 hour class" "done for the day" ? Probably two 1 hour classes in a row and nothing else for the day is quite tolerable so "reducing time required to stay on campus" might favor the class/class/done version. And probably class/class/class/class in a row would be pretty tiring, but how much better or worse would class/break/class/class/class be compared to class/class/break/class/class or class/break/class/class/break/class ?

In order to find the best schedule, you need formulas for all of the factors, like exhaustion compared to not having to stay on campus longer than needed compared to reducing number of times needing to travel. If you have two classes separated by a break, then what is the longest break that you would stay on campus to reduce travel, compared to going home (or to work or whatever) because the break is so long that you are just wasting time waiting around? Back when I went to university, if I had an 8:30 to 9:30 am class and then no second class until 13:30, I had a tendency to skip the early class because the 4 hour wait between classes did not seem worth getting up at like 6:30 in the morning -- how do you measure factors like that to come up with a way to decide which is the best of several schedules?

[yiwen yim]

Okay. Do you have any examples or samples suggestions that i could refer for this project? Thank you for all your help so far.

[Walter Roberson]

I do not have any examples.

The main thing is that you need to decide what it is you are minimizing and what factors you will take into account, and what your inputs are.

What are you calculating anyhow? Is this timetabling from the perspective of the institution, trying to allocate classrooms and number of teachers and making sure that for any given year and all programs that the required courses do not conflict with each other? Or is it timetabling from the perspective of the student where the course schedules are given and the student must pick between courses and sections ?

[yiwen yim]

okay. Noted. Thank you. I am trying to implement it from the perspective of institution which is to allocate the suitable classroom and the suitable lecturers and also making there will be no clashes as well as conflicts. thank you.