Using intelligent agents for self-adaptation and self-optimization of energy consumption in power networks

Global stabilization of energy networks without centralized control is the challenge this paper addresses. An agent-based approach to decentralized self-management of networked appliances is the solution this paper explores. Software agents represent thermostatically controlled appliances (TCAs), generate energy plans for expected energy consumption, and interact with each other as peers within a tree based peer-to-peer overlay. The Energy Plan Overlay Summation (EPOS) mechanism proposed, propagates plans generated by individual TCA agents to aggregators within this structure to achieve self-optimization/stabilization of energy requests. Preliminary results in a small-scale and restrictive environment are promising: a 15% increase in energy stabilization is achieved.