Application of neural network combined genetic algorithm to rank the development priority of heavy oil reservoirs

Abstract

Purpose

The purpose of this paper is to present a new approach for selecting the good heavy oil reservoirs to develop preferentially, which can avoid the huge economical loss resulted from wrong decision.

Design/methodology/approach

A new method of ranking the development priority of heavy oil reservoir is present, in which the neural network is applied for the first time to acquire reservoir parameters' weights through training samples and the genetic algorithm is used to optimize the joint weighs of neurons in case that neural network falling into local minimum. Additionally, the paper establishes subordinate function of every parameter. Eventually, comprehensive evaluation values of all heavy oil reservoirs are obtained.

Findings

The method can ensure the veracity and creditability of the parameters' weights, avoid the randomicity brought by experts.

Research limitations/implications

Accessibility of the data of many heavy oil reservoirs is the main limitation.

Practical implications

A very useful and new method for the decision makers of heavy oil reservoirs development.

Originality/value

The new approach of ranking the development priority of heavy oil reservoir based on the neural network and the genetic algorithm. The paper is aimed at the leaders who manage the development of heavy oil reservoirs.

Citation

Wu, X., Shi, J., Chen, F. and Wang, Y. (2009), "Application of neural network combined genetic algorithm to rank the development priority of heavy oil reservoirs", Kybernetes, Vol. 38 No. 10, pp. 1684-1692. https://doi.org/10.1108/03684920910994042