wseas-12204-20261013125433-1aa6ed 20261013125433232 wseas/wseas wseas.group@gmail.com WSEAS International Journal of Computational and Applied Mathematics & Computer Science 2769-2477 06 10 2026 06 10 2026 6 Anas Dahabiah Faculty of Information Technology Al-Sham Private University Damascus SYRIA Smaller language models often exhibit high confidence in erroneous predictions, undermining traditional ensemble methods. In heterogeneous ensembles, simple averaging can grant disproportionate influence to overconfident but incorrect outputs. This paper introduces a reliability-weighted extension of the Proportional Conflict Redistribution rule (PCR6), where model-specific weights are derived from validation accuracy. We provide a formal treatment of the conflict measure, prove that the weighted rule avoids Zadeh's paradox in the high-conflict regime, and establish a quasi-associativity bound of order O(K_max²). We further justify linear weighting from both Bayesian (conjugate Dirichlet) and minimax perspectives, and compare the approach with logarithmic pooling and Dirichlet calibration. Empirical evaluation on a multi-model MMLU setting shows that the primary practical benefit of weighted PCR6 in our experiments lies in improved calibration (substantially lower Expected Calibration Error) rather than large accuracy gains over simple averaging. Accuracy improvements appear most plausible when ensemble members differ markedly in reliability; when models are comparably strong, the method remains competitive while producing better-calibrated probabilities. The fusion step itself is computationally lightweight (tens of microseconds per query for small ensembles). 10 13 2026 10 13 2026 116 10 https://creativecommons.org/licenses/by/4.0/ 10.37394/232028.2026.6.10 https://wseas.com/journals/articles.php?id=12204 Achiam, J., et al. (2023). GPT-4 Technical Report. arXiv:2303.08774. Chen, L., et al. (2023). FrugalGPT. arXiv:2305.05176. Cobbe, K., et al. (2021). Training Verifiers to Solve Math Word Problems. arXiv:2110.14168. Dempster, A.P. (1967). Upper and Lower Probabilities Induced by a Multivalued Mapping. Ann. Math. Statist. 38(2):325–339. Dezert, J., & Smarandache, F. (2006). Proportional Conflict Redistribution Rules for Information Fusion. In Advances and Applications of DSmT, Vol. 2. Dubey, A., et al. (2024). The Llama 3 Herd of Models. arXiv:2407.21783. Hendrycks, D., et al. (2021). Measuring Massive Multitask Language Understanding. ICLR. Jiang, A.Q., et al. (2023). Mistral 7B. arXiv:2310.06825. Jiang, D., Ren, X., & Lin, B.Y. (2023). LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusion. ACL. Kuhn, L., et al. (2023). Semantic Uncertainty. ICLR. Lin, S., et al. (2022). TruthfulQA. ACL. Shafer, G. (1976). A Mathematical Theory of Evidence. Princeton University Press. Stanford Institute for Human-Centered Artificial Intelligence (HAI) (2025). The 2025 AI Index Report. Vectara (2025). Hallucination Evaluation Model Leaderboard. Wang, J., et al. (2024). Mixture-of-Agents. arXiv:2406.04692. Xiong, M., Hu, Z., Lu, X., Li, Y., Fu, J., He, J., & Hooi, B. (2024). Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMs. ICLR. Zadeh, L.A. (1979). On the Validity of Dempster's Rule of Combination. Memo M79/24, UC Berkeley. Kull, M., et al. (2019). Beyond temperature scaling: Obtaining well-calibrated multi-class probabilities with Dirichlet calibration. NeurIPS. Krogh, A., & Vedelsby, J. (1995). Neural Network Ensembles, Cross Validation, and Active Learning. Advances in Neural Information Processing Systems (NeurIPS). Zhang, P., Zeng, G., Wang, T., & Lu, W. (2024). TinyLlama: An Open-Source Small Language Model. arXiv:2401.02385.