Abstract
This paper presents a novel algorithm for solving a Student-Project Allocation problem, a variation of a two-sided matching algorithm, where a large number of students are allocated to numerous projects with limited available allocation places. Several new or improved algorithms are proposed, inspired by recommender systems and combined into a genetic algorithm. The main two algorithms are (i) variating the master list of student average grade ranking inspired by a fuzzy approach and (ii) extending the incomplete project preference lists by exploring the similarities of student choices, thus minimizing the number of randomly allocated students. The algorithm was implemented for allocating more than 500 students to more than 200 projects, in a course at the University of Zagreb, Faculty of Electrical Engineering and Computing. We compared our algorithm with the Deferred Acceptance (DA) Algorithm to check its validity. Using the fuzzy approach, the number of unallocated students decreased by 10%, with almost no effect to the top-ranked students. The combined usage of all proposed algorithms increased the number of successfully allocated students by more than 25%.
Highlights
Matching two sets of members, where each member has a preference over the members in other sets, has its application in various real-life problems
– RQ1: How can the number of unallocated students be reduced while retaining the importance of study grades in the master preference list?
THE SUMMARY OF FINDINGS After implementing and evaluating our algorithm, we propose the following answers to our research questions: RQ1: How can the number of unallocated students be reduced while retaining the importance of study grades in the master preference list?
Summary
Matching two sets of members, where each member has a preference over the members in other sets, has its application in various real-life problems. This paper describes a new algorithm for the StudentProject Allocation problem It applies to the Student-Project Allocation problem when the master list of ranked students is used. Students create incomplete preference lists, and the number of available places on projects is equal to the number of students and limited for each project This new setting creates a different environment than the one described by Gale-Shapley and requires a different approach. Two-sided matching algorithm with a preference list is used to allocate the students to the projects. A combination of priority in student allocation, incomplete students’ preference lists, and a limited number of places in projects produces a number of unallocated students. This paper proposes a new algorithm for solving the two-sided matching problem It is evaluated within the educational context to answer the following two research questions:. The last section concludes our paper with a discussion and provides future work suggestions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
