Incentive Mechanisms for Smart Grid: State of the Art, Challenges, Open Issues, Future Directions

A Survey on Soft Computing Techniques for Federated Learning- Applications, Challenges and Future Directions

Federated Learning is a distributed, privacy-preserving machine learning model that is gaining more attention these days. Federated Learning has a vast number of applications in different fields. While being more popular, it also suffers some drawbacks like high communication costs, privacy concerns and data management issues. In this survey, we define federated learning systems and analyse the system to ensure a smooth flow and to guide future research with the help of soft computing techniques. We undertake a complete review of aggregating federated learning systems with soft computing techniques. We also investigate the impacts of collaborating various nature-inspired techniques with federated learning to alleviate its flaws. Finally, this paper discusses the possible future developments of integrating federated learning and soft computing techniques.

Demand Response Contextual Remuneration of Prosumers with Distributed Storage

Prosumers are emerging in the power and energy market to provide load flexibility to smooth the use of distributed generation. The volatile behavior increases the production prediction complexity, and the demand side must take a step forward to participate in demand response events triggered by a community manager. If balance is achieved, the participants should be compensated for the discomfort caused. The authors in this paper propose a methodology to optimally manage a community, with a focus on the remuneration of community members for the provided flexibility. Four approaches were compared and evaluated, considering contextual tariffs. The obtained results show that it was possible to improve the fairness of the remuneration, which is an incentive and compensation for the loss of comfort. The single fair remuneration approach was more beneficial to the community manager, since the total remuneration was lower than the remaining approaches (163.81 m.u. in case study 3). From the prosumers' side, considering a clustering method was more advantageous, since higher remuneration was distributed for the flexibility provided (196.27 m.u. in case study 3).