Application of PSO (particle swarm optimization) and GA (genetic algorithm) techniques on demand estimation of oil in Iran
Introduction
The outlook of energy in Iran shows the importance of the need for systematic optimization of the energy use in Iran. Energy resources are limited and depleting [1]. Furthermore, Iran’s population is steadily increasing. The energy consumption in Iran is also rapidly increasing [2].
Iran, with population of more than 68 million [3], is one of the largest producers of crude oil in the world. In contrary to the public’s perception, Iran’s share of the market for high quality oil is as little as 2%. More specifically, while Iran has the fourth highest oil production rate, the oil produced in Iran is ranked 14th in terms of the quality [2].
Oil, natural gas, electric power, solar, wood, animal and plant waste are Iran’s primary energy sources [4].
Section snippets
Literature review
Several studies are presented to propose some models for energy demand policy management using different techniques. Unler developed PSO (Particle Swarm Optimization) energy demand models to estimate energy demand based on economic indicators in Turkey [5]. Canyurt and Ozturk presented Turkey’s fossil fuels demand estimation models by using the structure of the Turkish industry and economic conditions based on GA (Genetic Algorithm) [6]. Toksari developed Ant Colony energy demand estimation
GAs
The formulation of GAs was made about a decade after the first Evolutionary Strategies and Evolutionary Programming applications.
The theoretical basis of GAs lies in the concept of schema (plural schemata) [22]. Schemata represent solution templates where each location can be defined or left unspecified. The larger the number of uninstantiated location is, the greater the number of potential solutions that a schema represents. Schemata leading to higher fitness individuals are propagated
PSO
The PSO algorithm was first proposed by Eberhart and Kennedy [35], inspired by the natural flocking and swarming behavior of birds and insects. The concept of PSO gained in popularity due to its simplicity. Like other swarm-based techniques, PSO consists of a number of individuals refining their knowledge of the given search space. The individuals in a PSO have a position and a velocity and are denoted as particles. The PSO traditionally has no crossover between individuals, has no mutation and
Problem definition
Iran’s oil demand by using the structure of the Iran socio-economic conditions is the main objective of this study. GA and PSO demand estimation models (GA–DEM and PSO–DEM) are developed to estimate the future oil demand values based on population, GDP (gross domestic product), import and export figures.
The fitness function, F(x), takes the following form:where Eactual and Epredicted are the actual and predicted oil demand, respectively, m is the number of
Results and discussions
The PSO algorithm was coded with MATLAB 2007 and for GA algorithm, toolbox of MATLAB 2007 was used. The convergence of the objective function and sensitivity analysis were examined for varying user-specified parameters of GA (population size, methods of selection, reproduction, crossover, mutation and generation) and PSO (particle size (n), inertia weight () and maximum iteration number (t)). Each user-specified parameter combination (for each algorithm) was tested 10 times.
It was found that
Conclusion
Energy consumption growth rate shows the importance of the need for systematic optimization of the energy consumption in Iran.
Artificial intelligence methods have been successfully used to estimate Iran’s oil demand based on the structure of the Iran socio-economic conditions. 25 Years’ data (1981–2005) has been used for developing two forms (linear and exponential) of the GA and PSO demand estimation models. Three scenarios are designed in order to estimate Iran’s oil demand during 2006–2030.
Acknowledgments
The authors are grateful for the support provided for the present work by the Islamic Azad University, Dezful Branch and Iran’s Ministry of Energy. The authors also highly appreciate the constructive comments of the reviewers.
References (42)
- et al.
The crisis of gasoline consumption in the Iran’s transportation sector
Energy Policy
(2008) - et al.
Sustainability of energy production and use in Iran
Energy Policy
(2007) Sectoral energy consumption in Iran
Renewable and Sustainable Energy Reviews
(2005)Improvement of energy demand forecasts using swarm intelligence: the case of Turkey with projections to 2025
Energy Policy
(2008)- et al.
Application of genetic algorithm (GA) technique on demand estimation of fossil fuels in Turkey
Energy Policy
(2008) - et al.
A simulated-based neural network algorithm for forecasting electrical energy consumption in Iran
Energy Policy
(2008) - et al.
An integrated fuzzy regression algorithm for energy consumption estimation with non-stationary data: a case study of Iran
Energy
(2010) - et al.
Estimation of electricity demand of Iran using two heuristic algorithms
Energy Conversion and Management
(2010) - et al.
Forecasting the transport energy demand based on PLSR method in China
Energy
(2009) - et al.
The energy demand and the impact by fossil fuels use in the Mexico City Metropolitan Area, from 1988 to 2000
Energy
(2006)
Electricity demand for Sri Lanka: a time series analysis
Energy
An oil demand and supply model incorporating monetary policy
Energy
Development of future energy scenarios with intelligent algorithms: case of hydro in Turkey
Energy
Grey prediction with rolling mechanism for electricity demand forecasting of Turkey
Energy
Greek long-term energy consumption prediction using artificial neural networks
Energy
An analytical method for the measurement of energy system sustainability in urban areas
Energy
Forecast of oil reserves and production in Daqing oilfield of China
Energy
Short-term load forecasting of power systems by combination of wavelet transform and neuro-evolutionary algorithm
Energy
Electricity demand loads modeling using AutoRegressive Moving Average (ARMA) models
Energy
Reproduction in generational and steady state genetic algorithm
A comparison of selection schemes used in genetic algorithms
Cited by (198)
AI-based macro model learning for high cycle fatigue assessment of welded joints in large-span steel structures
2024, International Journal of FatigueCurrent status, future prediction and offset potential of fossil fuel CO<inf>2</inf> emissions in China
2023, Journal of Cleaner ProductionResearch on equipment corrosion diagnosis method and prediction model driven by data
2022, Process Safety and Environmental Protection