| Phase | 4-The |
| Prog. | PhD STI |
| Adm. – Grad. | 2017 – 2024 |
| Dir.; Codir. | Stéphane Gagnon |
| https://www.linkedin.com/in/miloudeloumri/ | |
| UQO | https://di.uqo.ca/id/eprint/1667/ |
Integrating Semantic Reasoning in a Multi-Agent System in Erlang
Eloumri, Miloud
Human Resources Management (HRM) is crucial for organizational success. It involves managing several key processes, including recruitment and selection. The recruitment process focuses on attracting suitable applicants, while the selection process involves choosing the best candidates based on organizational needs. These processes significantly contribute to gaining competitive advantages, but they face several challenges that can increase costs and complicate the task of selecting the best candidate.
This thesis uses Erlang functional programming language to enhance the efficiency of selecting the best applicants for a certain position. Erlang is known for its scalability, concurrency, reliability, real-time processing, and pattern matching capabilities, enabling it to effectively handle large and complex datasets.
However, Erlang’s application in personnel selection is largely unexplored. Therefore, the thesis aims to develop a sophisticated job-applicant matching and evaluation application that integrates the rule engine capabilities of SERESYE (Swarm-oriented ERlang Expert System Engine), the semantic wed data modeling functionality of Semantic Web Toolkit for Erlang Applications, and the principles of Multi-Criteria Decision Making (MCDM) using Weighted Sum Method (WSM) to enable comprehensive rule processing and multi-criteria evaluation of applicants qualifications against job requirements. Additionally, the thesis explores expanding the matching system’s data interpretation scope using RDFLib, a Python library for parsing most semantic data syntaxes.
The study implements a prototype use case represented by a simple ontology in the domain of relatives to explore the concept of rule matching over semantic web data using the mentioned Erlang technologies. However, the research primarily applies the matching approach to a complex area represented by Business Technology Management (BTM) jobs and applicants ontology modeled using Protégé ontology editor. The findings demonstrate the feasibility of integrating SERESYE rule processing with semantic web datasets, leading to enhanced job-applicant matching. This thesis thus makes a significant contribution to personnel selection, expert systems, semantic web, and decision-making systems, providing a foundation for future advancements.
Future work includes addressing the limitations of current implementation. The development of a web interface for the job-applicant matching system using Erlang’s Zotonic framework. The use of SERESYE multiple parallel rule engines and Erlang Poolboy library for creating and managing a pool of concurrent process. The application extensions to include related operations such as emailing, scheduling interviews, and making job offers.