Multi-Agent Framework for Spatial Load Forecasting

Abstract

A multi-agent framework for spatial electric load forecasting, especially suited to simulate the different dynamics involved on distribution systems, is presented. The service zone is divided into several sub-zones, each subzone is considered as an independent agent identified with a corresponding load level, and their relationships with the neighbor zones are represented as development probabilities. With this setting, different kind of agents can be developed to simulate the growth pattern of the loads in distribution systems. This paper presents two different kinds of agents to simulate different situations, presenting some promissory results.