Decentralized energy management platform

1. A method for power management, comprising:

applying a decentralized control to manage a large-scale community-level energy system;

obtaining a global optimal solution satisfying constraints between agents representing the energy system's devices as a state-based potential game with a multi-agent framework;

independently optimizing each agent's output power while considering operational constraints and assuring a pure Nash equilibrium (NE), wherein a state space helps coordinating the agents' behavior in energy system to deal with system-wide constraints including supply demand balance, battery charging power constraint and satisfy system-wide and device-level (local) operational constraints; and

controlling distributed generations (DGs) and storage devices using the agent's output.

2. The method of claim 1 , comprising adding devices as agents into an energy system framework and submitting the agent's services or request from other agents, wherein agents dynamically join or leave the energy system at any time without interrupting the rest of the energy system.

3. The method of claim 1 , comprising handling supplier agents or devices to provide power to supply a load and submit their service as a supplier.

4. The method of claim 3 , comprising playing as an active agent in the multi-agent framework, wherein the agents include grid, diesel, PV, or battery.

5. The method of claim 1 , comprising handling a demand agent, wherein each load in a network submit its service as demand in the multi-agent framework and inform supplier agents about its requested load value.

6. The method of claim 1 , comprising applying a distributed dispatching strategy for energy sources and loads to minimize a total marginal cost of operation for the energy system and satisfy device-level and system-wide level operational constraints.

7. The method of claim 6 , wherein the dispatching strategy comprises applying a game theory, a state-based potential game, a learning algorithm, a gradient play or a Newton method.

8. The method of claim 1 , comprising determining a total cost of operation for the energy system (system-wide cost model) as a summation of the devices' operational cost (local cost models).

9. The method of claim 1 , wherein system-wide operational constraints are satisfied through passing message packages and negotiations between the agents and agent local constraints are handled locally by a related agent.

10. The method of claim 1 , wherein the agents use one or more learning algorithms at each iteration of the game.

11. A method for power management, comprising:

applying a decentralized management platform based on multiagent framework and state-based potential game;

analytically obtaining a global optimal solution satisfying constraints between agents (energy system's devices);

applying the power management as a game where each agent is able to independently optimize its output power while considering its operational constraints;

assuring a Nash equilibrium (NE) exists.

12. The method of claim 11 , comprising using a state-based potential game where a state space added to the game helps with coordinating the agents' behavior in energy system to deal with system-wide constraints including supply demand balance, battery charging power constraint, wherein the NE of the game is efficient with respect to system total cost of operation and satisfies system-wide and device-level (local) operational constraints.