Proving Regulatory Compliance: Full Compliance Against an Expressive Unconditional Obligation is coNP-Complete

TL;DR

This paper investigates the computational complexity of proving regulatory compliance, demonstrating that full compliance against an expressive unconditional obligation is coNP-complete, and maps the complexity of various related problem variants.

Contribution

It establishes the coNP-completeness of proving full compliance for expressive unconditional obligations and maps the complexity landscape of related compliance verification variants.

Findings

Proving full compliance against an expressive unconditional obligation is coNP-complete.

The complexity of verifying compliance varies across different obligation variants.

The paper provides a comprehensive complexity map for these compliance verification problems.

Abstract

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.