A study on a practical consequence finding system

SOLAR是一个一阶推论发现系统。为了提高系统的可扩展性，提出了SOL表演算的分治策略，开发了一种新的基于最佳优先搜索的结果发现系统，并介绍了系统的各种剪枝技术。对于命题案例，我们开发了一个基于SAT技术的结果发现系统原型。新的一阶太阳系与旧的一阶太阳系相比表现出更好的性能，所开发的SAT解算器GlueMiniSat用作命题系统的推理引擎，分别在2011年和2013年的应用UNSAT竞赛中获得第一和第二名。

SOLAR is a first-order consequence finding system. To improve the scalability, we proposed a divide-and-conquer strategy for SOL tableau calculus, developed a new consequence finding system based on best-first search, and introduced various kinds of the pruning techniques for the system. For propositional cases, we developed a prototype of a consequence finding system based on SAT technologies. The new first-order SOLAR system showed superior performance compared to the old one, and the developed SAT solver, called GlueMiniSat, which is used as a inference engine of the propositional system, got 1st and 2nd places of SAT 2011 and 2013 competitions in Applications UNSAT category, respectively.