Agent-Based Spatial Load Forecasting

Long-term load forecasting aims at predicting the evolution of the electric consumption in a certain area in order to resize the grid in accordance. There are two components to study: the increment in existing consumption and the appearance of new clients. We focus here on the latter. With this purpose, we present an ongoing work that applies agent-based modelling to this end. Representing each existing electric customer by an agent, candidate agents will decide their settlement based on their inquiries to present agents on their likes and actual urban constraints obtained from real sources (e.g. prohibition to build upon a lot marked as a park). We will test this system with a Monte-Carlo simulation and compare the obtained results with real data to validate our approach.