Yuquan Shan

Generation bidding game with potentially false attestation of flexible demand

With the onset of large numbers of energy-flexible appliances, in particular plug-in electric and hybrid-electric vehicles, a significant portion of electricity demand will be somewhat flexible and accordingly may be responsive to changes in electricity prices. In the future, this increased degree of demand flexibility (and the onset of only short-term predictable intermittent renewable supply) will considerably exceed present level of uncertainty in day-ahead prediction of assumed inelastic demand. For such a responsive demand idealized, we consider a deregulated wholesale day-ahead electricity marketplace wherein bids by generators (or energy traders) are determined through a Nash equilibrium via a common clearing price (i.e., no location marginality). This model assumes the independent system operator (ISO) helps the generators to understand how to change their bids to improve their net revenue based on a model of demand-response. The model of demand-response (equivalently, demand-side bidding day ahead) is based on information from load-serving entities regarding their price-flexible demand. We numerically explore how collusion between generators and loads can manipulate this market. The objective is to learn how to deter such collusion, e.g., how to set penalties for significant differences between stated and actual demand, resulting in higher energy prices that benefit certain generators.

Multiperiod subscription pricing for cellular wireless entrants

We consider a two-player game involving a large incumbent (or incumbent oligopoly) and small entrant into a cellular-wireless access provider marketplace. The entrants customers must pay roaming charges. We assume that the roaming charges are transparent to the user and regulated to prevent an incumbent from creating barriers to entry in the marketplace. To be able to reckon suitable (regulated) roaming charges, in this paper we consider a potentially stricter model of competition than [7] (though still not all subscribers to the lowest-cost provider), and a revenue function for the entrant that considers future revenue streams when its deployment is greater and its customers therefore do not roam as much, i.e., a multiperiod/longitudinal revenue model.

A simulation framework for uneconomic virtual bidding in day-ahead electricity markets: Short talk

About two thirds of electricity consumers in the United States are served by Regional Transmission Organizations (RTOs) and Independent System Operators (ISOs). One of their primary responsibilities is the operation of organized auctions for purchasing and selling electricity that have a two-settlement structure with coordinated day-ahead (DA) and real-time (RT) energy markets. The DA market takes place on the day before the actual power dispatch, and creates a financial obligation to deliver and withdraw power from the transmission grid. In contrast, the RT energy market is a physical market where predicted and actual supply and demand of electricity are balanced on the delivery day. Purely financial transactions, known as virtual bids, were introduced in wholesale electricity markets to allow participants (including energy traders that do not control generation assets or serve load) to exploit arbitrage opportunities arising from expected price differences between day-ahead and real-time energy markets and to enhance convergence between DA and RT prices. More specifically, virtual demand (supply) bids are financial positions for the purchase (sale) of energy in the DA market, which are settled with a countervailing offer to sell (buy) at the RT price without the bidder taking title to physical electricity. Virtual demand bids are typically referred to as DECs, while virtual supply bids are known as INCs. Virtual bids clear with generation and load bids in the DA market, and may set the DA market-clearing price. Virtual bids have strong interactions with other elements of the electricity market design. For instance, Financial Transmission Rights (FTRs) are financial contracts to hedge transmission congestion between two nodes in the transmission network (a source and a sink defined in the contract), and entitle their holders the right to collect a payment when day-ahead congestion arises between the source and the sink [1]. Since FTRs settle at the day-ahead prices, virtual bids could be placed in the day-ahead energy market in order to affect day-ahead electricity prices in a direction that enhances the value of the FTRs. In our study, we consider a model of the DA electricity market at any node in the network. Market participants include power generators and loads submitting physical bids, and financial players placing virtual bids. Virtual bids affect the DA market clearing prices, but we assume that they have no impact on RT prices. Theoretical results on interior Nash equilibria are given, assuming that virtual bidders can perfectly predict RT prices and hold no FTRs [2] sinking at the node. We then adopt a hypergame framework [3] to model the DA market, assuming imperfect prediction of RT prices by different virtual bidders. When no market participant holds FTRs, virtual bidders help achieve convergence between DA and RT nodal prices, as expected [4]. In this setting, we also allow one virtual bidder to hold a FTR position sinking at the node. Our numerical results show that, with FTR as another source of revenue, the larger the FTR position, the greater the incentive for the FTR holder to place uneconomic virtual bids at the FTR sink to enhance the value of her financial position, in line with [5, 6]. We also show that the manipulation causes not only losses for other virtual bidders, but also the divergence between DA and RT prices. Methods for detecting such uneconomic bidding are also investigated. Our technical report is available at http://www.cse.psu.edu/research/publications/tech-reports/2016/CSE-16-003.pdf.

Multicommodity games in public-cloud markets considering subadditive resource demands

In still developing, public cloud-computing markets, prices for virtual machine (VM) offerings fluctuate, and not just for spot/preemptible instances. Moreover, some (particularly derivative) providers allow for fine-grain initial resource provisioning and dynamic reprovisioning of VMs. In this preliminary study, we consider long-lived tenants of a public cloud under resource-based service-level agreements. For noncooperative aggregative multicommodity (plural IT resource) games among them, conditions are established for an interior Nash equilibrium and an exact potential (giving convergence to Nash equilibrium). Also discussed are extensions to plural IT resource demands that are subadditive in workload intensity.

A simulation framework for uneconomic virtual bidding in day-ahead electricity markets

Virtual bids were introduced in U.S. wholesale electricity markets to exploit arbitrage opportunities arising from expected price differences between day-ahead and real-time energy markets. These financial instruments have interactions with other elements of the electricity market design. For instance, virtual bids may be intended to move day-ahead electricity prices in a direction that enhances the value of Financial Transmission Rights (FTRs) settling at those energy prices. We consider a model of the day-ahead electricity market at one node in the network, under the assumption that virtual bidding does not affect the real-time dispatch of generators. Theoretical results on interior Nash equilibria are presented, assuming virtual bidders can perfectly predict real-time prices and hold no FTRs. We then adopt a kind of hypergame framework to model the day-ahead market, assuming imperfect prediction of real-time prices by different virtual bidders, and present simulation results with and without FTRs. Finally, we discuss two detection mechanisms that may be used by regulators to distinguish between competitive and manipulative market outcomes, as well as trade-offs between specificity and sensitivity.

Optimal Power Flow with Random Wind Resources

We consider optimal power-flow calculations based on statistics (means and variances) of wind-power that are assumed known hour ahead. In real-time, adjustments of these calculations to the true wind power amounts have been proposed using affine control and chance constraints on generation and transmission. We study how correlated wind power sources affect the optimized power plan and cost. We also explore how the errors in wind forecasting statistics, due to real-time control actions to limit wind oscillations (ramping) or just simple misinformation, affect costs. There are direct extensions to the case where random/flexible demands/loads are correlated and similarly predictable.