基于模型诊断方法是为了克服传统模拟方法的缺陷而兴起的智能技术，对人工智能领域的理论研究和集成电路产业发展具有重要推动作用。然而，基于模型诊断方法和技术无法满足当前超大规模集成电路发展的需要，诊断过程已成为集成电路工业设计流程的瓶颈。. 现有结合MaxSAT的诊断方法未能有效利用模型中蕴涵等诸多结构信息。本项目拟从模型的结构和行为出发，挖掘拓扑、统治和蕴涵等结构信息，给出融合结构信息的MaxSAT诊断方法，以显著提高诊断求解效率。对于大规模诊断问题，构建融合问题结构信息和枚举树特征的并行求解诊断完备方法，研究结构信息指导格局检测和布尔约束传播等优化策略的随机求解诊断不完备方法，以提升可处理诊断问题的规模。最后，研制原型系统对本项目提出的方法进行检验。. 项目的预期成果将丰富和发展基于模型诊断的理论与方法，显著提高其实用性。

Model-based diagnosis method is a newly-raised intelligent technology to overcome the defects of the traditional ones, which plays strong promotion effect on the theoretical research in the field of artificial intelligence as well as the prosperity of integrated circuit industry. However, model-based diagnosis methods and its relevant techniques cannot satisfy the demand of super-scale integrated circuits. More importantly, the diagnosis process has become a bottleneck of the integrated circuit industry design. . Recently, combined with Maxsat, some diagnostic methods fail to effectively utilize the structure information in the model, such as implication, which starts from the structure and the behavior of the model and then explores the structure information such as topology, dominance and implication, etc. In this project, we propose the MaxSAT diagnosis method combined with structure information to effectively improve the efficiency. For solving large-scale problem diagnosis, we build a complete approach to the diagnosis of parallelism based on the structure information and the characteristic of the enumeration tree, and consider the incomplete stochastic diagnosis method combined with optimization strategies such as structure information guiding configuration checking and Boolean constraint propagation in order to enhance the scale of the diagnosis problem. Furthermore, a prototype system is built up to check our methods.. The achievements of this project are expected to enrich and develop the theory and method of model-based diagnosis as well as significantly improve its practicability.