Abstract
This study aims to formulate and solve the natural gas storage facility valuation problem by emphasizing the impact of observable gas prices on storage operational policies. In a deregulated market, non-energy players could enter the market and rent the storage facilities. As market prices of gas are revealed, the leaseholders dynamically adjust their operational decision to derive the maximum benefit from the price fluctuation. However, the direct influence of available gas prices on storage values has been neglected in the literature. The current work derives the market filtration based on observable asset prices, and formulates the storage problem within the Brody-Hughston-Macrina (BHM) X-factor pricing theory to fill the gap in the literature. The essence of this research is to use the stochastic dynamics of price revelation in back-testing procedure rather than to calibrate the spot and forward prices. This effort theoretically contributes to improve the BHM pricing framework for randomly timed cash-flow assets and to construct the market filtration and σ-algebra incorporating two uncertainty sources of price fluctuation and random switch times. Thereafter, a numerical scheme is developed based on Monte Carlo regression to solve the corresponding optimal switching problem. Then, this Least Square Monte Carlo (LSM) technique is extended to the Info-LSM method in order to back-test the resulted strategy on the available information. The designed decision-making algorithms are efficient because they possess polynomial time and memory complexities. In terms of performance, the developed LSM algorithm reduces the total execution time by approximately 50%, especially in the case of a storage problem with a linear ordinary differential equation inventory level. Moreover, the experiments depict the empirical convergence of both schemes to the exact solutions with smaller standard deviations. Furthermore, the extended technique provides a very short list of the admissible strategies and indicates that the majority of simulated strategies are not acceptable, subject to the current market situations. In conclusion, this study enables decision-makers to solve the switching problems and back-test the solution on the observable asset prices without using any financial instruments.
Original language | English |
---|---|
Place of Publication | Universiti Teknologi Malaysia |
Publisher | |
Publication status | Published - 2 Jun 2014 |