Scalable Minimization Algorithm for Partial Bisimulation

TL;DR

This paper introduces a scalable algorithm for computing partial bisimulation, a relation between simulation and bisimulation, which improves efficiency in minimizing labeled transition systems with large state spaces.

Contribution

The authors present a novel minimization algorithm and tool that efficiently handles partial bisimulation by scaling with the subset of bisimulated actions, addressing prior scalability issues.

Findings

Algorithm scales with the size of the bisimulated action subset

Tool efficiently computes partial bisimulation for large systems

Improves over existing minimization methods for simulation and bisimulation

Abstract

We present an efficient algorithm for computing the partial bisimulation preorder and equivalence for labeled transitions systems. The partial bisimulation preorder lies between simulation and bisimulation, as only a part of the set of actions is bisimulated, whereas the rest of the actions are simulated. Computing quotients for simulation equivalence is more expensive than for bisimulation equivalence, as for simulation one has to account for the so-called little brothers, which represent classes of states that can simulate other classes. It is known that in the absence of little brother states, (partial bi)simulation and bisimulation coincide, but still the complexity of existing minimization algorithms for simulation and bisimulation does not scale. Therefore, we developed a minimization algorithm and an accompanying tool that scales with respect to the bisimulated action subset.