ISEP - DM – Engenharia de Inteligência Artificial
URI permanente para esta coleção:
Navegar
Percorrer ISEP - DM – Engenharia de Inteligência Artificial por assunto "Agent communities"
- A novel multi-level and community-based agent ecosystem to support customers dynamic decision-making in smart gridsPublication . Pereira, Helder Jorge Vilarinho; Vale, Zita Maria Almeida doElectrical systems have evolved at a fast pace over the past years, particularly in response to the current environmental and climate challenges. Consequently, the European Union and the United Nations have encouraged the development of a more sustainable energy strategy. This strategy triggered a paradigm shift in energy consumption and production, which becoming increasingly distributed, resulted in the development and emergence of smart energy grids. Multi-agent systems are one of the most widely used artificial intelligence concepts in smart grids. Both multi-agent systems and smart grids are distributed, so there is correspondence between the used technology and the network's complex reality. Due to the wide variety of multi-agent systems applied to smart grids, which typically have very specific goals, the ability to model the network as a whole may be compromised, as communication between systems is typically non-existent. This dissertation, therefore, proposes an agent-based ecosystem to model smart grids in which different agent-based systems can coexist. This dissertation aims to conceive, implement, test, and validate a new agent-based ecosystem, entitled A4SG (agent-based ecosystem for smart grids modelling), which combines the concepts of multi-agent systems and agent communities to enable the modelling and representation of smart grids and the entities that compose them. The proposed ecosystem employs an innovative methodology for managing static or dynamic interactions present in smart grids. The creation of a solution that allows the integration of existing systems into an ecosystem, enables the representation of smart grids in a realistic and comprehensive manner. A4SG integrates several functionalities that support the ecosystem's management, also conceived, implemented, tested, and validated in this dissertation. Two mobility functionalities are proposed: one that allows agents to move between physical machines and another that allows "virtual" mobility, where agents move between agent communities to improve the context for the achievement of their objectives. In order to prevent an agent from becoming overloaded, a novel functionality is proposed to enable the creation of agents that function as extensions of the main agent (i.e., branch agents), allowing the distribution of objectives among the various extensions of the main agent. Several case studies, which test the proposed services and functionalities individually and the ecosystem as a whole, were used to test and validate the proposed solution. These case studies were conducted in realistic contexts using data from multiple sources, including energy communities. The results indicate that the used methodologies can increase participation in demand response events, increasing the fitting between consumers and aggregators from 12 % to 69 %, and improve the strategies used in energy transaction markets, allowing an energy community of 50 customers to save 77.0 EUR per week.