A Cooperative Ant Colony Optimization-Genetic Algorithm approach for construction of energy demand forecasting knowledge-based expert systems

Knowledge-based expert systems are becoming one of the major tools for scientists and engineers nowadays, since they have many attractive features and can be called upon to deal with real/complex engineering application problems which are not easy to solve by orthodox methods. Meanwhile, increasing worldwide demand for different types of energy requires development of advanced intelligent forecasting tools to provide a basis from which decisions and plans can be made. This study presents a new approach called Cooperative Ant Colony Optimization-Genetic Algorithm (COR-ACO-GA), to construct expert systems with the ability to model and simulate fluctuations of energy demand under the influence of related factors. The proposed approach has two main stages, at the first stage it uses genetic algorithms to generate data base of the expert system, and at the second stage it adopts ant colony optimization to learn linguistic fuzzy rules such that degree of cooperation between data base and rule base increases and consequently performance of the algorithm improves. We evaluate capability of COR-ACO-GA by applying it on three case studies of annual electricity demand, natural gas demand and oil products demand in Iran. Results indicate that COR-ACO-GA provides more accurate-stable results than adaptive neuro-fuzzy inference systems (ANFISs) and artificial neural networks (ANNs), and can assist decision makers in making appropriate decisions and plans for a coming period. (c) 2012 Elsevier B.V. All rights reserved.