As a classical aggregation tool, the weighted average method iswidely used in information fusion. It is the Lebesgue integral withrespect to the weights essentially. Due to some inherent interactionamong diverse information sources, the weighted average method doesnot work well in many real problems. To describe the interaction, anintuitive and effective way is to replace the additive weights with anonadditive set function defined on the power set of the set of allinformation sources. Instead of the weighted average method, weshould use the Choquet integral or some other nonlinear integrals,especially, the new nonlinear integral introduced by the authorsrecently. The crux of making such an improvement is how to determinethe nonadditive set function from given input-output data when thenonlinear integral is viewed as a multiinput single-output system. Inthis paper, we employ a specially designed genetic algorithm torealize the optimization in determining the nonadditive set function.
展开▼
