Skill-Based Group Allocation of Students for Project-Based Learning Courses Using Genetic Algorithm: Weightless Penalty Model

摘要

In software engineering courses, project-based learning (PBL) is an essential counterpart for assessment. PBL requires a good spread of students based on their individual skills, this pushes for a successful outcome. Traditionally, the students are assigned to the group randomly by facilitators. In group assignment problem (GAP) students need to be placed in the appropriate groups encapsulating on general and specific criterions where possible to provide diversity. However, the need for a system arises where the assignment of students to groups require not only students to be based on certain criteria but also takes into account specific constraints. Constraints allow taking parameter such as skill preference in an unevenly or limited distributed skill set. For successful completion of projects, there is a need for groups that share same strength. In this paper, a generic method is proposed that uses the genetic algorithm to generate evenly balanced groups. We have employed weightless penalty function to rank the preference of certain constraints based on skills and incur a penalty if they are not satisfied. Since the benchmark datasets are unavailable, data collected from software engineering courses of our University is made available and its utilization with the proposed method is shown.