Power system demand management contract design: A comparison between game theory and artificial neural networks

Demand management in power systems is one of the many innovative ideas introduced recently for the efficient management of our burgeoning energy needs. The major aim of demand management programs is to control an ever increasing demand for electric power among customers. Controlling customers demand for electric power will bring about load relief for electric utilities and an overall improvement of power system security. There exists a wide variety of demand management programs. Recently demand management contracts have been designed in which incentives are offered to customers who willingly sign up for load interruption. In this scenario the incentive offered to the customer should be greater than the cost of interruption while simultaneously making it beneficial for the utility. Mechanism design from Game theory has hitherto been used in the design of such contracts. This paper proposes a novel idea where an artificial neural network is designed to select the optimal contract. Our designed artificial neural model uses the back propagation learning algorithm with useful system parameters serving as the input and the output the value of the contract. The results obtained from Game theory using mechanism design are used to benchmark the results obtained from this artificial neural network. Furthermore our results show that artificial neural networks are an attractive option for contract design as they require less computational procedures than game theory without much of a compromise on accuracy.