An Embeddable Self-Reflective Self-Adaptive Software Model for Resource Management

Abstract

With the growing complexity of today's world, there is rise in the design of smart and autonomous technologies. These technologies (or systems) are usually software-intensive and should have ability to self-manage, optimize resources, eliminate or reduce human intervention and still achieve quality and optimal output(s). Thus, technologies of this nature should self-reflect and self-adapt to internal and external factors to achieve its desired goal. Self-reflective, selfadaptive concepts are (or can be) used in a number of different areas including: robotics systems, autonomous computing, space systems, manufacturing, recommender systems, and e-commerce (for transactions monitoring), among others. These technologies basically have software component, which are usually underpinned by model(s). The main goal of this research, therefore, is to develop an embeddable software model (with self-reflective, selfadaptive properties) that shows real-time resource management. In developing our model, we particularly took into cognizance MAPE-K template for identification of self-* properties and performance requirements. We applied a Case Study within a particular domain (the education sector) to demonstrate the usability of this software model in managing human resource (a primary resource). The software model is able to reflect and modify its behavior at run-time in response to internal and external factors. Evaluation results show that the software model successfully self-manage the deployment of human resources; it was able to determine over/under-staffing, and redistribute (and/or identify recruitment needs for) human resource within the domain under consideration.