Fairness Pricing in the Energy Market
Stahl, Chad L.
MetadataShow full item record
With the current payment structure of the energy market, consumers are billed based on the kilowatt hours (kWh) that a consumer uses during some time frame, generally a month. This research uses existing framework in customer fairness pricing, modified to account for the energy usage predictability of a customer in the energy market. The goal of this research is to establish an efficient and useful algorithm for rewarding customers based on their predictability; providing incentives for more predictable use and disincentives for poor predictability, as well as the subsequent results which would be of importance to an energy service provider. We will primarily focus on existing research into fairness pricing and Distribution Cost Sharing and briefly look into Willingness-To-Pay (WTP), Cost Allocation Problems and Traveling Salesman Games (TSGs). Various models which are relevant to fairness pricing are presented, with research into the Distribution Cost Sharing problem serving as the core base of the research developed in this thesis. Several cost allocation papers which aided in formulation of both the Distribution Cost Sharing problem and this research are presented, primarily discussing the proportional willingness-to-pay model and cooperative cost allocation in Traveling Salesman Games. The models are effective at billing customers fairly based on their proportional service cost, taking into account factors such as single distance and customer density while ensuring all customers are served. The research in this thesis differs from these existing models by pricing customer's based on their predictability as opposed to the aforementioned criteria. Elements of the Distribution Cost Sharing problem are studied to create the fairness aspects of the customer predictability model developed in this thesis. The model presented thus closely resembles a typical Distribution Cost Sharing problem, with the objective of maximizing the revenue of the system while proportionally yet fairly pricing customers based on some criteria. The result is a model which aims to maintain the desired or existing revenue for energy service providers, while ensuring that customers are priced fairly based upon their contribution to the excess cost present in the energy market which are typically passed along to all customers through higher utility prices. A computational study on the model developed in this thesis was performed using several quasi-randomly generated data sets, with three different sized sets to test model run-times on smaller versus larger sets. Our results reveal that in regards to clustering customers the developed model presents a fairer and individualized pricing scheme through the predictability clustering, than through a simple flat-rate payment scheme in which all customers pay for the cost of unpredictability. In regards to model run-time, the developed model performed well with smaller models while some of the larger data sets could take significantly longer. Comparison graphs are displayed in this thesis which demonstrate the effectiveness of the model for both the price clustering and run-time.