An ExpTime Procedure for Description Logic $\mathcal{ALCQI}$ (Draft)

Yu Ding

DOI: https://doi.org/10.48550/arXiv.cs/0703051

2007-03-19

Abstract:A worst-case ExpTime tableau-based decision procedure is outlined for the satisfiability problem in $\mathcal{ALCQI}$ w.r.t. general axioms.

Logic in Computer Science

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to determine the concept satisfiability problem in Description Logic (DL) ALCQI. Specifically, the author aims to propose a tableau - based decision procedure to determine whether a concept in ALCQI is satisfiable in the presence of general inclusion axioms.

### Problem Background

1. **Description Logic ALCQI**: This is an extended description logic that supports qualified number restrictions, that is, it can express statements such as "There are at least n objects in the R - relation that satisfy the condition C".
2. **Concept Satisfiability Problem**: Given a concept formula and a set of terminological axioms, determine whether there exists an interpretation such that the concept formula is true.
3. **Complexity**: This problem belongs to the ExpTime complexity class, which means that in the worst - case, exponential time is required to solve it.

### Main Challenges

- **Qualified Number Restrictions**: Traditional C - rules (such as Ramsey's Rule) are not applicable to description logics with qualified number restrictions.
- **Inverse Roles**: When dealing with inverse roles, traditional tableau methods may encounter difficulties because inverse roles introduce additional constraints.

### Solutions

The author proposes a new method to solve the above problems through the following steps:

1. **Introducing Cut Formulae**:
- For each modal subformula, introduce cut formulae to simplify the reasoning process.
- These cut formulae allow inverse roles to be regarded as independent role names, thus avoiding the complexity brought by inverse roles.

2. **Fine - Tuning Modal Constraints**:
- Adjust the cardinality constraints in the modal constraints of successor nodes according to the choice of predecessor nodes.
- This can handle qualified number restrictions more precisely.

3. **Linear Integer Inequalities (LIIs)**:
- Use a system of linear integer inequalities to represent and solve modal constraints.
- Generate all possible role - filler combinations through atomic decomposition techniques and convert them into a system of linear integer inequalities.

4. **Tableaux Expansion Rules**:
- Define a set of tableaux expansion rules (such as PB - rule and LII - rule) for constructing and expanding the tableaux tree.
- These rules ensure the correctness and completeness of the tableaux tree.

5. **Termination Guarantee**:
- Use a restart strategy and a blocking mechanism to ensure that the algorithm terminates within a finite time.
- Each restart will derive a new inconsistency set, thus ensuring that an answer can be finally found.

### Conclusion

This paper successfully proposes a tableau - based decision procedure with ExpTime complexity in the worst - case for solving the concept satisfiability problem in description logic ALCQI. This method not only deals with the complexity of qualified number restrictions and inverse roles, but also ensures the correctness and termination of the algorithm.

### Formula Examples

- **Modal Constraints**:
\[
(\exists \leq n R.C)^\mathcal{I} = \{ x \in \Delta^\mathcal{I} \mid | \{ y \in \Delta^\mathcal{I} : (x, y) \in R^\mathcal{I} \text{ and } y \in C^\mathcal{I}\} \leq n \}
\]