SimSG: Rule-based Simulation using Stochastic Graph Transformation.

Abstract

Stochastic models can be found in various domains. For example, biochemical processes such as molecular interactions or the dynamics of wireless network topologies, where changes occur with certain probabilities. Having the ability to simulate scenarios in these domains can be crucial when real-life observations of certain processes are infeasible, e.g., protein-protein interactions in biochemistry, or expensive, e.g., building large wireless networks for research purposes. Stochastic graph transformation systems provide the means to describe the structure and simulate the behavior of such probability-driven environments in an adequate way, by modelling the state transitions using graph transformation rules, whose application depends on the current state and their application probabilities. To the best of our knowledge, there is currently no general-purpose simulation tool available anymore that performs rule-based simulations using stochastic graph transformation. Therefore, we developed SimSG a modular stochastic simulation tool that addresses the needs of a wide range of application domains-in contrast to most specialized simulation tools that are limited to one domain only. To facilitate the versatility of the tool, SimSG can be configured to employ different general-purpose tools for incremental graph pattern matching (currently, Democles and Viatra). We evaluate SimSG based on two use cases: First, using an example of the biochemistry domain, we conduct a comparative evaluation against the domain-specific tool KaSim. Second, we underpin the general-purpose applicability of SimSG by analyzing the simulation of a wireless sensor network scenario.

Stochastic Graph Transformation For Social Network Modeling

“…In this paper we use SimSG [12] to simulate rule-based models, such as the basic voter model, for a given start graph and rates according to the semantics of stochastic graph transformations [16]. SimSG implements its own version of Gillespie's well-known algorithm [14] which describes the behavior of stochastic systems over time using a continuous-time Markov Chain with exponentially distributed transition delays.…”

“…In this section we simulate these models with the CTMC-based SimSG tool [12] to see how, with suitably chosen parameters, their behavior matches that of the original formulation [11]. In particular, we want to answer the following questions.…”

“…Finally, in view of simulation experiments, we would like to advocate the use of dedicated and highperformance Gillespie-style stochastic graph transformation software such as in particular SimSG [12] (cf. Section 4).…”

“…In the second pair of rules, an agent connected to one with a different opinion adopts the opinion of the other. We could now associate rates with these rules, resulting in a stochastic graph transformation system that allows simulation using the SimSG tool [12] and the derivation of differential equations using the rule algebra approach [8,2,6,7].…”

A Simulator for Probabilistic Timed Graph Transformation Systems with Complex Large-Scale Topologies

“…A link between GTSs and discrete event simulation has been considered in [14]. Also, an extension of GTSs with stochastic behavior and related simulators like GraSS [15] and SimSG [4] have been developed. However, to the best of our knowledge, no simulator for GTS variants that support timed and probabilistic behavior (like PTGTSs [9]) has been presented so far.…”