An Evolutionary Strategy with a Tailor-Made Repair Operator for Natural-Gas Management Under Supply Shortages

Abstract

Managing natural gas (NG) demand under uncertainty and limited storage capacity is especially critical to determine sustainable strategies that lead to significant savings. In this paper an evolutionary approach is proposed to support policy decision-making when NG supply shortages are imposed. The eventual winter scarcity of NG in countries like Argentina, China, South Korea, India, Canada and Australia, where industrial plants are forced to operate under low-supply conditions, has become a crucial budget problem. Evolutionary algorithms based on Darwinian principles have proved to be promising to handle real-world problems whose treatment involves complex and nonlinear optimization. Hence, the strategy is based on a Real Coded Genetic Algorithm, which was enhanced with specialized repair operators. A dynamic chromosome contemplates the long-term horizons, thus contributing to the accuracy of realistic results. Since the decoded solution may sometimes be infeasible, a specially designed procedure is always applied to repair it. The proposed methodology allows to systematically address changes in feedstock supply and market demand as a whole. Efficient tactical planning was achieved by optimizing the NG amounts provided. This strategy could locate feasible convenient solutions within reasonable computing times.