Fault Diagnosis and Optimization for Agent Based on the D-S Evidence Theory

摘要

To solve only consider the evidence oneself in fault diagnosis conflict using Dempster-Shafer evidence theory(D-S), not consider environment influence and the different capacities of diagnosis method, and sometimes because of the more subjectivity, the more qualitative factor and the less quantitative analysis, the fairness of tender evaluation is suspected. The fault diagnosis and optimization for Agent based on the D-S evidence theory is proposed. Firstly, the dynamical adjustment of Agent weight which is integrated into the D-S classified optimization and Agent with rewards and punishments mechanism as the main content is introduced. Secondly, the weight is constantly corrected according to the Agent diagnosis result to avoid the subjectivity and form a closed loop using the adjustment weight, the optimization result and environment feedback. Finally, the test result shows that our proposed method can raise the accuracy rates of diagnosis and improve optimization precision and ensure algorithm reliability.