TOOL DEVELOPMENT FOR UNIVERSITY COURSE TIMETABLING PROBLEM CASE STUDY: ITB UNIVERSITY COURSE TIMETABLING
University timetabling in ITB generally submitted by each Department (Prodi). But for first <br /> <br /> <br /> level course (TPB) and all public lecture scheduled by Directorate of Education (Dirdik). This <br /> <br /> <br /> can make a problem because many...
Saved in:
Main Author: | |
---|---|
Format: | Final Project |
Language: | Indonesia |
Online Access: | https://digilib.itb.ac.id/gdl/view/22922 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Institut Teknologi Bandung |
Language: | Indonesia |
Summary: | University timetabling in ITB generally submitted by each Department (Prodi). But for first <br />
<br />
<br />
level course (TPB) and all public lecture scheduled by Directorate of Education (Dirdik). This <br />
<br />
<br />
can make a problem because many course clashed because it is regulated by each Department. <br />
<br />
<br />
Also the schedule are not evenly distributed where some students with dense schedule on a <br />
<br />
<br />
certain day and empty the other days. <br />
<br />
<br />
Timetabling algorithm that will be implemented is greedy algorthm and genetic algorithm. Both <br />
<br />
<br />
algorthm was optimization algorithm that often used for searching solutions for any problem. All <br />
<br />
<br />
problem from ITB university course timetabling will be analyzed to be constraint for each <br />
<br />
<br />
algorthm while finding the solution. <br />
<br />
<br />
The result from this final project is that algorthm can make schedule with optimal whre have <br />
<br />
<br />
least constraint violation. The average of vitness value form greedy algorithm result is 0.967. <br />
<br />
<br />
While the genetic algorthm result in 100th <br />
<br />
<br />
, 300th, and 500th generation is 0.986, 0.913 and 0.939. <br />
<br />
<br />
The result from combining two algorthms where initiation of genetic used result from greedy <br />
<br />
<br />
algorithm is more optimal and get better result. The average fitness value from 100th generation <br />
<br />
<br />
is 0.986. <br />
<br />
<br />
From the result, the most constraint violation in the solution is where course that have some <br />
<br />
<br />
course number with same Departement participant must scheduled at the same time. This is <br />
<br />
<br />
because the lack of available room. Adding the available room will make the better result from <br />
<br />
<br />
the tool. |
---|