1 Hit in 0.036 sec

An Intelligent Service-Based Layered Architecture for eLearning and eAssessment

Qaisar Javaid, Muhammad Arif, Shahnawaz Talpur, Umair Ahmed Korai, Munam Ali Shah
2017 Mehran University Research Journal of Engineering and Technology  
The rapid advancement in ICT (Information & Communication Technology) is causing a paradigm shift in eLearning domain. Traditional eLearning systems suffer from certain shortcomings like tight coupling of system components, lack of personalization, flexibility, and scalability and performance issues. This study aims at addressing these challenges through an MAS (Multi Agent System) based multi-layer architecture supported by web services. The foremost objective of this study is to enhance
more » ... is to enhance learning process efficiency by provision of flexibility features for learning and assessment processes. Proposed architecture consists of two sub-system namely eLearning and eAssesssment. This architecture comprises of five distinct layers for each sub-system, with active agents responsible for miscellaneous tasks including content handling, updating, resource optimization, load handling and provision of customized environments for learners and instructors. Our proposed architecture aims at establishment of a facilitation level to learners as well as instructors for convenient acquisition and dissemination of knowledge. Personalization features like customized environments, personalized content retrieval and recommendations, adaptive assessment and reduced response time, are believed to significantly enhance learning and tutoring experience. In essence characteristics like intelligence, personalization, interactivity, usability, laidback accessibility and security, signify aptness of proposed architecture for improving conventional learning and assessment processes. Finally we have evaluated our proposed architecture by means of analytical comparison and survey considering certain quality attributes. An Intelligent Service-Based Layered Architecture for eLearning and eAssessment the college students are online students [2]. E-Learning systems are leveraging latest technologies like APIs, Cloud, Web services, HTML5 and MAS to increase overall utility and efficiency of these systems. AI concepts and methodologies are being adopted to develop intelligent, personalized, reliable and fault-tolerant systems for eLearning. Traditional eLearning systems suffer from significant shortcomings i.e. scalability, personalization, interoperability, coupling and performance issues [3-5]. Newer technologies like web services, cloud, semantic web, xAPIs and advanced methods like IR, KE, service matchmaking, and Query optimization can significantly contribute towards overcoming these issues. Since last decade, intelligent agent-based eLearning systems have received significant attention of research community. Agent based technologies can facilitate learning processes in numerous ways including learnersystem interaction, personalization, AI model generation for learning and simulation, resource optimization and a common infrastructure to absorb diverse range of software components [6-7]. Agent-based information system provides direct instructions or feed back to the students without human intervention [8]. Software agents as part of an MAS, need to communicate with other agents through a message passing mechanism [9]. Internal state and behaviors of member agents are not accessible by other member agents of MAS [10-11]. Agents can only discover and invoke services offered by other agents by means of middleware services [10,12]. These middleware services are provided by MAS platform in which agents are embedded. Member agents can interact using ACL (Agent Communication Language) while communication protocols define the rules of interaction among member agents of MAS [11]. However, this communication is not always local and homogeneous, as remote interaction with member agents of other MAS systems and non-agent software components is quite likely condition [10]. Interoperability between heterogeneous member agents and heterogeneous MAS is considered as a major research challenge which has received significant attention from research community [12-18]. We have proposed an MAS-based multi-layered architecture supported by Web services. This architecture mainly comprises of two modules namely eLearning and eAssessment. Five constituent layers of eLearning module ensure the loose coupling and distributed nature of the system. These layers include Human interface, resource, MAS, dB controller, and DSRs (Data Storage Repositories) layers. eAssessment architecture also consists of five layers namely; Human interface, resource, FDL (Functional Description Layer), implementation and database layer. Detailed description of these modules is provided in subsequent sections. Section 2 provides related work in areas of MAS, SOA, personalized leaning and query processing. Section 3 provides general architecture with detailed description of the proposed system and underlying layers. Section 4 provides evaluations mechanism followed by conclusion in section 5. RELATED WORD MAS is a sub-discipline of DAI (Distributed Artificial Intelligence). MAS deals with diverse range of operations including distribution or decomposition of activities, communication, coordination and conflict resolution among agents [19]. A large number of MAS based solutions for eLearning can be seen in existing literature. An architecture by [20] has utilized agents for each subject including expression extraction and user interface. ACM CR classification hierarchy is used in this regard. In [21] authors introduced extension to MASHA (Multi Agent System Handling Adaptively) to MASHA-EL aimed at supporting eLearning. In this system students An Intelligent Service-Based Layered Architecture for eLearning and eAssessment
doi:10.22581/muet1982.1701.10 fatcat:4uy75gh6gjcajpfqmsqti35x6a