Implementation of a RESTful API-Based Evolutionary Algorithm in a Microservices Architecture for Course Timetabling
Year: 2024
Type: article
Abstract: Introduction/Main Objectives: Implement an evolutionary algorithm within a RESTful API for a course timetabling system that employs a microservices architecture. Background Problems: The current course timetabling at Politeknik Statistika STIS uses the third-party application (aSc Timetables), which lacks a generator as a service, resulting in its inefficiency due to the lack of integration with SIPADU NG. Novelty: The evolutionary algorithm is built as a service (RESTful API) within a microservices architecture and supports custom constraints for timetables. Research Methods: One of the evolutionary algorithm families, the (1+1) evolutionary strategy, is implemented and used to create a course timetable 1000 times. Each course timetable created will have its cost calculated to assess the goodness of the algorithm implementation. The developed RESTful API is also evaluated through black box testing. Finding/Results: For the odd semester data, 40.5% of the trials yielded a cost value between 4 and 5, while for the even semester, all trials produced a cost value below 1. The resulting cost value is close to 0, which indicates that the timetable created has minimal violations. Additionally, black box testing concluded that the service operates as expected, delivering the anticipated output.
Authors: Zuhdi Ali Hisyam, Farid Ridho, Arbi Setiyawan
Institutions: Politeknik Statistika STIS, Jiangsu University
Language: English
Field: Decision Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: diamond
OpenAlex ID: W4405761132