Organisations are required to show that their procedures and processes satisfy the relevant regulatory requirements. The computational complexity of proving regulatory compliance is known to be generally hard. However, for some of its simpler variants the computational complexity is still unknown. We focus on the eight variants of the problem that can be identified by the following binary properties: whether the requirements consists of one or multiple obligations, whether the obligations are conditional or always in force, and whether only propositional literals or formulae can be used to describe the obligations. This paper in particular shows that proving full compliance of a model against a single unconditional obligation whose elements can be described using formulae is coNP-complete. Finally we show how this result allows to fully map the computational complexity of these variants for proving full and non compliance, while for partial compliance the complexity result of one of the variants is still missing.
翻译：要求各组织表明其程序和过程符合相关的监管要求。证明遵守监管要求的计算复杂性一般已知是困难的。然而,对于其一些较简单的变式来说,计算复杂性仍然未知。我们侧重于以下二元属性可以查明的问题的八个变式:要求是否包括一项义务或多项义务,是有条件的还是始终有效的,以及是否只能用假设的字面或公式来描述这些义务。本文件特别表明,证明对单一无条件义务的模型完全遵守,而该模型的要素可以用公式来描述。最后,我们表明,这一结果如何使这些变式的计算复杂性能够完全描绘出来,以证明完全和不遵守,而对于部分遵守来说,一个变式的复杂结果仍然缺失。
相关内容
微信扫码咨询专知VIP会员