Towards a Negotiation-Based Multi-Agent Power Management System for Electric Vehicles

Increasingly complex load requirements of electric vehicles (EV) can be fundamentally divided into propulsion and non-propulsion loads. Further segregation of the non-propulsion loads into multi-priority, multi-time constant electrical burdens presents a compelling case for a real world negotiation environment in which the negotiation issues are meeting the power requirements under constrained power resources. This paper discusses the formulation of a negotiation framework, in which agents can learn and adapt when making deals on behalf of classified EV loads to meet their power requirements. The agents can use genetic algorithms to maximize the payoff in a highly dynamic negotiation environment according to certain attributes. Simulation results are presented to show that the proposed scheme offers a promising framework towards ongoing investigations into intelligent agent-based vehicular power and energy management schemes.