Electricity Swing Option Pricing by Stochastic Bilevel Optimization: a Survey and New Approaches

Abstract

We demonstrate how the problem of determining the ask price for electricityswing options can be considered as a stochastic bilevel program with asymmetricinformation. Unlike as for financial options, there is no way for basingthe pricing method on no-arbitrage arguments. Two main situations are analyzed:if the seller has strong market power he/she might be able to maximizehis/her utility, while in fully competitive situations he/she will just look for aprice which makes profit and has acceptable risk. In both cases the seller hasto consider the decision problem of a potential buyer - the valuation problemof determining a fair value for a specific option contract - and anticipate thebuyer's optimal reaction to any proposed strike price. We also discuss somemethods for finding numerical solutions of stochastic bilevel problems witha special emphasis on using duality gap penalizations.