An evidence-combination-based simulation result validation method for multi-source data

Abstract

Simulation result validation is a crucial work in simulation credibility assessment. The result validation metric is a measure of agreement between simulation output and experimental observations. Sometimes the observations maybe derive from different data sources such as the actual system output, credible hardware-in-the-loop simulation system output and the expert opinions and so on. Then the agreement analysis results of system response would be multiple certainly. To solve the simulation result validation with multi-source data, the paper proposes a result validation method based on evidence combination. First, the evidence representation methods for multi-source data on the structure of evidence space are proposed. Then the multiple evidence bodies are aggregated based on evidence combination rules and the integrated validation result could be achieved. In the end, the application process and effectiveness of this validation method is illustrated through a numerical example.