The energy system is highly complex and its performance depends on many decision makers, such as consumers, producers of raw materials, refineries, government bodies, etc. Each decision maker has its own objectives and constraints, which can be conflicting, such as choosing whether to fund research to lower production costs or building capacity expansions or coincident, such as securing the necessary amount of energy. Furthermore, dwindling hydrocarbon reserves, energy security and the implications of energy consumption on the environment have all driven the development of a number of energy technologies which are on the cusp of gaining ubiquitous use. Long-term research and infrastructure planning on energy systems can only be made with an understanding of the process by which these technologies will be incorporated into the existing energy system.
Agent-based modeling is a growing area with successful applications in manufacturing, consumer markets, supply chains, etc [1][2][3]. Although there is not a universal definition for agents, in general an agent may be defined as an identity that makes independent decisions, interacts with its environment, has a defined goal, is autonomous and learns. Multi-agent systems are particularly adept at showcasing the interactions between individual entities in complex systems and the aggregate system behaviors that result from these interactions. Hence, multi-agent systems are suitable for analyzing the energy system and they have already been applied to electricity market analysis [4].
In this work, we propose a robust and scalable multi-agent framework where characteristics of energy systems constituents and their interactions can be configured to study the system behavior. The framework is composed of six (6) major agent classes. The Raw Material Producer Agent class comprises the extraction or growth of raw materials (coal, oil, corn, etc.) while the Producer Agent class handles the production of finished products (gasoline, ethanol, electricity, etc.). The Producer Agents interact with the Consumer Agent class through the sale of products for consumption. Consumption from the residential, commercial, industrial and transportation sectors is considered. The Government Agent class can influence the actions of the producer and consumer agents through the use of indirect methods such as taxes or subsidies or through the direct support for research on a particular technology. This research is conducted by Research Agents that make decisions on which technologies seem to be most promising and then submit proposals for work which will improve the current status of a technology. Finally, the Environment Agent is responsible for the accurate portrayal of the effects of the system agents on the “state of the world” outside the system boundaries and vice versa. Therefore, the interactions between system components are not fully specified, since each player in the energy system is modeled as an intelligent agent capable of making independent decisions. Allowing the system interactions to be determined by the agents themselves can produce simulated behavior which mirrors realistic behavior that may not be readily apparent from the behavior of each individual.
A particular focus of the model is the role of research in the development and adoption of new technologies. Each technology modeled follows a learning curve representing the work necessary to bring it to a state of production readiness. Once at the production stage, an experience curve is used to help represent the challenges associated with the technology adoption process.
After applying the proposed agent-based simulation framework to the energy system of the state of Indiana, we are then able to analyze how different entities within the system react to both internal and external stimulus. Results will be presented on the effects of both government subsidies on emerging technologies as well as taxes on existing technologies. In addition, the effects of new technologies in both the emergent and production stages on the use of current technologies will be explored.
References:
[1] A.I. Anosike, D.Z. Zhang, “An agent-based approach for integrating manufacturing operations,” International Journal of Production Economics (In Press)
[2] E. Guerci, S. Ivaldi, S. Pastore, S Cincotti, “Modeling and implementation of an artificial electricity market using agent-based technology,” Physica A: Statistical Mechanics and its Applications, pp 69-76 Volume 355 Issue 1 (2005)
[3] N. Julka, R. Srinivasan, I. Karimi, “Agent-based supply chain management-1: framework,” Computers & Chemical Engineering, pp 1755-1769 Volume 26 Issue 12 (2002)
[4] V. Koritarov, “Real-World Market Representation with Agents: Modeling the Electricity Market as a Complex Adaptive System with an Agent-Based Approach,” IEEE Power and Energy Magazine, pp. 39-46, July/August 2004