Agent-based approach for Decentralized Genetic Algorithm

Abstract

The progressive development of local energy communities as well as the growing interesting in Machine Learning approaches for the energy sector are leading to deep changes in to Power Systems sector. In this evolution scenario, the concept of local energy markets would lead to new operations and management approaches. It is well-established that markets based on distributed ledgers are suitable to the system requirements rather than centralized approaches. This study aims at introducing an agent-based model for simulating participants' offering mechanism. In this way the market users' behaviours are simulated in a realistic manner. This paper improves and enhances the results obtained from previous authors study, where a decentralized version of the common genetic algorithm was presented. The proposed agent-based framework has been implemented on a Real Time Digital Simulator, capable of inspect the grid dynamics. Finally, in order to reproduce participants behaviour, IoT smart devices was implemented.