Heuristic Techniques for Improved Problem-Solving Strategies

本研究的最终目的是了解 分析启发式信息的性质， 改进问题解决策略的启发式技术。 该项目旨在扩大和评估 通过识别易处理类的约束求解算法 问题，并改善一般回溯计划， 另一 我们的方法是将 其拓扑结构的问题。 此外，我们将扩大 表示非单调知识的约束网络模型 引导真理维护算法 这一扩展将 应用于诊断和时间推理领域。

The ultimate objectives of this research are to understand and analyze the nature of heuristic information and to develop heuristic techniques for improved problem solving strategies. This project seeks to extend and assess the efficiency of constraint-solving algorithms by identifying classes of tractable problems on one hand and improving general backtrack schemes on the other. Our approach is to relate the tractability of the problem to its topological structure. In addition we will extend the constraint-network model to represent non-monotonic knowledge and guide Truth-maintenance algorithms. This extension will be applied to the domains of diagnosis and temporal reasoning.