MIKADO - Gestão inteligente de energia em tempo real: modelos baseados em dados e em conhecimento para tomada de decisão utilizando semântica e aprendizagem automática

The MIKADO Project aims to take advantage of the integration of advanced Machine Learning (ML) and Knowledge-Based (KB) methodologies to contribute for the evolution of Automated Energy Management (AEM). Positioned at the forefront of interdisciplinary research, the project combines expertise from Power and Energy Systems and the domain of Artificial Intelligence, notably ML, to address unexplored challenges within the energy sector. MIKADO's comprehensive strategy is designed to conceiving, developing, and implementing advanced models, methods, and tools, namely user-centric AEM solutions to a broad spectrum of stakeholders, including energy professionals, policymakers, and end-users, thereby navigating the complexities inherent in contemporary energy management. To achieve its objectives, MIKADO has delineated a set of ambitious goals: • Development of an Advanced Adaptive Decision-Making Framework: MIKADO aims to synthesize ML and KB approaches to forge a highly adaptive decision-making architecture. This framework is intended to seamlessly integrate data-driven insights with heuristic reasoning, thereby ensuring optimal energy management decisions that are responsive to the dynamic and multifaceted nature of real-world scenarios. • Enhancement of Consumer Empowerment and Trust: A pivotal aspect of MIKADO's mission is to provide AEM systems that are not only efficacious but also intuitive and transparent for users across varying levels of technical expertise. By prioritizing explainability and user engagement, the project endeavors to demystify energy management technologies, fostering a higher degree of trust and participation among consumers in sustainable energy initiatives. • Promotion of Real-Time, Contextually Adaptive AEM Solutions: MIKADO is committed to extending the capabilities of current AEM technologies by ensuring the provision of solutions that are both agile and tailored to the immediate needs of the energy ecosystem. Leveraging real-time analytics and adaptive learning algorithms, MIKADO aims to equip AEM systems with the ability to dynamically adjust to evolving environmental conditions, user preferences, and emergent contextual challenges. Incorporating these strategic goals, MIKADO is conceived to develop a solution that dynamically discerns the most effective decision-making strategy for energy management in any given context, from the perspective of the consumer. This solution will be seamlessly integrated with smart grid and electricity market frameworks and technological platforms. By crafting a system that not only accounts for the present context and available decision-making time but also learns to prioritize rapid reasoning over conventional ML approaches, MIKADO will address scenarios often neglected by standard ML models. The project will utilize validation and evaluation metrics to continuously refine its algorithms for each specific context. Moreover, the explanations generated by the system will enhance user interaction, enrich the AI system's learning process, and aid in clarifying discrepancies between anticipated and actual model outcomes.