Evaluation of gas sales agreements with indexation using tree and least-squares Monte Carlo methods on graphics processing units

A typical gas sales agreement, also called a gas swing contract, is an agreement between a supplier and purchaser for the delivery of variable daily quantities of gas between specified minimum and maximum daily limits. The primary constraint of such agreements that makes them difficult to value is that the strike price is set based on the indexation principle, under which the strike price is called the index. Each month, the value of the index is determined by the weighted average price of certain energy products (e.g. crude oil) in the previous month. We propose a lattice-based method (trinomial trees) and a simulation-based method (least-squares Monte Carlo simulations) for pricing such swing contracts with indexation. With the help of graphics processing unit (GPU) technology, we can efficiently evaluate the algorithms. We also provide a detailed analysis using several numerical examples of the indexation and how different model parameters will affect both the optimal value and optimal decisions.