The central open question in Descriptive Complexity is whether there is a logic that characterizes deterministic polynomial time (PTIME) on relational structures. Towards this goal, we define a logic that is obtained from first-order logic with fixed points, FO(FP), by a series of transformations that include restricting logical connectives and adding a dynamic version of Hilbert's Choice operator Epsilon. The formalism can be viewed, simultaneously, as an algebra of binary relations and as a linear-time modal dynamic logic, where algebraic expressions describing ``proofs'' or ``programs'' appear inside the modalities. We show how counting, reachability and ``mixed'' examples (that include linear equations modulo two) are axiomatized in the logic, and how an arbitrary PTIME Turing machine can be encoded. For each fixed Choice function, the data complexity of model checking is in PTIME. However, there can be exponentially many such functions. A crucial question is under what syntactic conditions on algebraic terms checking just one Choice function is sufficient. Answering this question requires a study of symmetries among computations. This paper sets mathematical foundations towards such a study via algebraic and automata-theoretic techniques.
翻译:描述性复杂度的核心问题是,在关系结构上,是否存在确定性多元时间(PTIME)的逻辑特征。为实现这一目标,我们定义了从固定点的第一阶逻辑(FO(FP),通过一系列转换,包括限制逻辑连接,并添加Hilbert选择操作员Epsilon的动态版本。形式主义可以同时被视为二进制关系的代数和线性模式动态逻辑,其中代数表达“校准”或“方案”的代数出现在模式中。我们展示了计算、可达性和“混合”的例子(包括线性方程式模L2)是如何在逻辑中解析的,以及如何将任意的 PTIME 图灵机编码的。对于每个固定选择功能,模型检查的数据复杂性在 PTIME 中都是一个指数化的。但是,一个关键问题是,对于“校准”“校准”或“程序”的代数条件,在模型的数学基础中,我们展示了如何计算、可达标度和“混合”例子(包括线性方方程式)是如何在逻辑中进行辨测算的。