Optimal data fusion based on information quality function

Information quality based on entropy is defined to measure uncertainty of a probability distribution, which is an important criterion in the field of data fusion. For given probability distributions, how to obtain the reasonable result of high quality is still an open issue. Especially when given probability distributions are highly conflicted, the result of applying classic data fusion method is counter-intuitive. In this paper, an optimal data fusion method on the basis of genetic algorithm is proposed to resolve this issue and the result with minimal Gini entropy within a reasonable range can be obtained. This method can handle the probabilistic fusion problem of given sources high conflict effectively and the Gini entropy of result is relatively small. Genetic algorithm is applied in this method to handle data fusion problem in this paper. The effectiveness of this proposed method is illustrated and the advantages of this method are analysed compared with other fusion methods.