A Multi‐Objective Optimization and Decision‐Making Framework for Safety‐Critical Ground Vehicle Endurance Tests

Accelerated endurance testing for safety‐critical ground vehicles requires reproducing field‐induced fatigue within limited time while maintaining stable and reliable evaluation conditions. However, conventional procedures rely on empirical rules or component‐level metrics, offering limited capability for balanced and system‐consistent fatigue reproduction. This study presents an integrated, data‐driven framework combining Multi‐Objective Optimization (MOO) and Multi‐Criteria Decision Making (MCDM) to systematically support endurance test design for safety‐critical ground vehicles. For this, an NSGA‐II formulation was used to identify terrain–direction–speed combinations that satisfy mileage requirements while keeping each component within acceptable fatigue ranges. Subsequently, the CRITIC method derived objective criterion weights, and the non‐compensatory PROMETHEE II approach was used to rank Pareto‐optimal solutions to ensure robust and stable rankings. Prediction accuracy and loading consistency were evaluated using the Geometric Mean Relative Absolute Error (GMRAE) and the Factor of Damage Ratio (FDR). The framework successfully reproduced target fatigue damage, yielding optimal solutions with GMRAE values below 1.0 for most component groups and extreme channel rates below 3%. Crucially, the CRITIC‐PROMETHEE II integration demonstrated stable rank consistency under 10% and 20% weight perturbations, confirming its reliability and suitability for safety‐critical systems. This methodology converts empirical endurance scheduling into a quantitative, reproducible, and systematically balanced test design process applicable to a wide class of ground vehicles.