Forecasting Crude Oil Prices and Determining the Optimal Production Level Using the Evolutionary Pattern of Neural Networks and Nash Equilibrium

Being able to correctly predict oil price and production behaviour can help decision makers to adopt more appropriate policies to better regulate the provision of oil as a critical commodity in World trade. Due to the fluctuating and non-linear trend of supply and demand for crude oil and its price, smart and non-linear methods, especially evolutionary patterns based on neural networks are expected to have good predictive power for short-term crude oil prices. This paper applies the Neural Network colonial competition algorithm to evaluate oil prices for the period January 1982 to October 2015 using panel data for OPEC crude oil production and OECD oil consumtion for the period. We can compare this with optimal levels of production and consumption obtained using game theory and Nash equilibrium. We observe a Correlation Coefficient of R= 0.921104, confirming the explanatory power of the colonial competition algorithm. We further find that Neural networks output and game theory and Nash equilibrium can predict the optimal level of OPEC production and consumption of OECD countries for short periods of a month.