OptiRica: Towards an Efficient Optimizing Horn Solver

TL;DR

This paper introduces OptiRica, a framework for optimizing Horn solvers using a lattice-based approach, with a DSL for modeling and solving complex optimization problems like MaxSAT/SMT, demonstrated through network repair applications.

Contribution

It presents a novel Scala-based DSL framework that enables systematic modeling and optimization of Horn solving problems, extending existing solvers with optimization capabilities.

Findings

Framework successfully models various optimization problems.

Application to software-defined network repair demonstrates practical utility.

Potential to improve efficiency of Horn solvers in complex scenarios.

Abstract

This paper describes an ongoing effort to develop an optimizing version of the Eldarica Horn solver. The work starts from the observation that many kinds of optimization problems, and in particular the MaxSAT/SMT problem, can be seen as search problems on lattices. The paper presents a Scala library providing a domain-specific language (DSL) to uniformly model optimization problems of this kind, by defining, manipulating, and systematically exploring lattices with associated objective functions. The framework can be instantiated to obtain an optimizing Horn solver. As an illustration, the application of an optimizing solver for repairing software-defined networks is described.