Abstract:

Aiming at the critical challenge of small-sample modeling in ammunition consumption prediction for armored vehicle target damage assessment, progressive four-zone damage dynamic modeling is proposed. A hierarchical damage propagation mechanism is established through deconstruction of target structural trees. By integrating damage area ratios with efficacy weighting coefficients, this work pioneers dynamic cumulative rules with downward-inclusive efficacy, effectively overcoming limitations of traditional static damage threshold approaches. Gamma distribution characterization of target efficacy degradation is revealed. A bidirectional parameter correction model is developed by incorporating zonal damage probabilities, enabling precise characterization of ammunition consumption distribution patterns. Bayesian-Markov chain-Monte Carlo hybrid solving algorithm is developed. A prior information consistency verification conversion theorem is proposed, coupled with Jeffreys-simulation joint prior distribution construction and adaptive Metropolis-Hastings sampling strategies, achieving significant reduction in parameter estimation errors. Case validation demonstrates that the proposed framework maintains prediction errors below 5% across multiple damage severity levels, establishing a battlefield-adaptive analytical paradigm for small-sample damage assessment scenarios. This methodology advances conventional approaches through its integration of structural dynamics, statistical pattern recognition, and adaptive Bayesian computation, providing a robust theoretical foundation and practical toolkit for ammunition logistics optimization in modern warfare.

Key words: ammunition consumption, effectiveness destruction, Gamma distribution, armored vehicle, Bayesian inference, Markov chain-Monte Carlo