Electricity contract pricing and portfolio management

Given a scenario tree for uncertain electricity spot market prices and inflow to own reservoirs, we propose a method based on nonlinear programming to find an equivalent martingale measure in every node in the tree. The constraints in the program ensures absence of arbitrage opportunities, and the objective function ensures that the resulting equivalent martingale measure is as close as possible to a reasonable equilibrium measure reflecting the preferences of contract market participants. The method is applied to an electricity portfolio management problem using multistage stochastic programming, in which prices of electricity contracts need to be known in every node of the scenario tree.