证据推理理论近似方法研究

摘要

Dempster-Shafer (DS)和Dezert-Smarandache (DSm)证据推理理论是信息融合的有力工具。针对在证据处理过程中证据数量多以及证据组合时焦元爆炸使计算量迅速增大的问题，对证据与焦元控制方法、规则近似方法和蒙特卡洛(Monte Carlo)方法三种方法在DS理论和DSm理论下进行了分类分析，指出了下一步值得研究的三个问题。这为证据推理理论不同近似方法的使用提供了理论参考。%Dempster-Shafer (DS) and Dezert-Smarandache (DSm) evidence reasoning theories are powerful tools for information fusion. Aimed at the problem that the evidence number is more in the procedure of evidence processing and the explosion of focal elements enlarges the computation during the combination of evidences, this paper implements classification and analysis of those three methods including evidence and focal element control, rule approximation and Monte Carlo methods under the DS and DSm theories, and puts forward three problems to be studied next, which offers a theoretical reference for the use of diverse approximate methods of evidence reasoning theory.