DEPARTMENT SELECTION USING GAME THEORY
In Institut Teknologi Bandung (ITB), new students are not directly accepted into their departments, but will go through a preparation phase (TPB) in each faculties first for their first year. At the end of TPB, a department selection will be held to determine each student’s university major. Depa...
Saved in:
Main Author: | |
---|---|
Format: | Final Project |
Language: | Indonesia |
Online Access: | https://digilib.itb.ac.id/gdl/view/83750 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Institut Teknologi Bandung |
Language: | Indonesia |
Summary: | In Institut Teknologi Bandung (ITB), new students are not directly accepted into their
departments, but will go through a preparation phase (TPB) in each faculties first for
their first year. At the end of TPB, a department selection will be held to determine
each student’s university major. Department selection can be modeled as a game using
game theory, and the ideality of the selection result can be seen through this model.
The players in this game are the Faculty of Mathematics and Natural Sciences (FMIPA)
departments, with cooperative behavior. Each department has its own payoff function
that is associated with the correlation value between students and university majors,
based on the selection criteria. The ideality of the selection result can be seen through
each department’s payoff value, where a selection result is said to be ideal if the sum of
all departments’ payoff value is maximal.
The selection game can be modeled as a weighted complete bipartite graph, with the
partitions of the graph correspond with the students set and the departments set. The
weight of each edge corresponds with the correlation value between the student and
the department that is connected through the edge. This graph can be modified by
duplicating the department nodes so that the number of node each department has is
equal to the maximum quota of the department, then adding student nodes so that the
number of student nodes is equal to the number of department nodes. A maximumweight
perfect matching can be found on the modified graph using Hungarian algorithm.
The matching that is obtained from the algorithm represents the ideal selection result. |
---|