Published Online:September 2024
Product Name:The IUP Journal of Supply Chain Management
Product Type:Article
Product Code:IJSCM020924
Author Name:Ramesh Murthy and Padmalatha N A
Availability:YES
Subject/Domain:Strategic
Download Format:PDF
Pages:15
Accurate forecasting of petroleum product demand is crucial for economic planning and policy formulation. Various methods, ranging from simple moving averages to complex fuzzy logic, have been employed, yielding varying levels of accuracy. Recent machine learning approaches rely primarily on historical data for forecasting future demand This study argues that external variables have a significant effect on petroleum product demand. Key independent variables influencing the demand for LPG, petrol, jet fuel, kerosene, and diesel in India were identified. Using demand data and independent variable values from 1998 to 2023, individual regression models were constructed for each petroleum product. Initial models exhibited multicollinearity and autocorrelation, which were mitigated using ridge regression. The models’ predictive accuracy was evaluated using MAPE, SMAPE, and RMSE, yielding an average error rate of less than 10%. To predict the demand by March 2030, the best-case and worst-case scenario values of the independent variables were determined. The impacts of Indian government energy policies, technological developments, and geopolitical factors were also included. Based on these inputs, the demand by March 2030 was determined. The study forecasts that in the best case scenario, the demand for petroleum products by March 2030 will be 188.3 million tons, and in the worst case scenario, it will be 278 million tons.
Petroleum products are essential inputs for the social and economic development of any nation. India, as the world’s third-largest consumer of petroleum products, has experienced significant growth in petroleum consumption, with a compound annual growth rate (CAGR) of 4.36% from 2010 to 2020, surpassing 210 million metric tons