Combined approach architecture development to simulation modeling of systems with parallelism

Abstract

Paradigms and graphical-analytical tools for building simulation tools and forming the architecture of a combined approach to studying the dynamic properties of systems with parallelism are described. An extension of the formal language of Petri nets is presented, which has greater modeling power than WF nets. The properties of hierarchical Petri nets are used to synthesize a holistic model. Discrete-event modeling and modeling of dynamic systems, which allow reflecting the quantitative and qualitative characteristics of the elements of the systems under study, served as the basis for the combined approach to the simulation of systems with parallelism. On their basis, graphic-analytical tools are proposed that provide the ability to describe the modeled system, adhering to the principle of structural similarity. They have dynamic simulations that make it easy to visually analyze and correct the model. Also, the proposed toolkit provides for the analysis of the dynamic properties of the model, which makes it possible to identify accumulated phenomena that can lead to unpredictability of the system’s functioning. A conceptual model for the synthesis and analysis of systems with parallelism is proposed, which provides for the construction of the components of the model based on the architecture. Their step-by-step analysis and the formation of an integral model of the software system are carried out using a network representation, according to the matrix description of which invariants are calculated. The analysis of invariants allows one to obtain the dynamic properties of the model and determine the localization of structures that lead to critical situations when they are detected. The architecture of the combined approach to the simulation of systems with parallelism is built, which provides the study of their dynamic properties to improve the reliability of the functioning of software systems

Improving the efficiency of dynamic analysis of the combined simulation model for software with parallelism

“…It is proposed to use a combined approach to simulation modeling of systems with parallelism [12]. But its linear use does not always lead to the desired result.…”

“…In order to increase the effectiveness of the combined approach to simulation modeling of software with parallelism [12], it is necessary to take into account the peculiarities of the construction and dynamic analysis of simulation models of software components when developing the method's for the synthesis and analysis of a combined simulation model of a software system with parallelism. Also, the method's should take into account the peculiarities of the study of static and dynamic properties of components on artificially closed models.…”

“…-the principle of structural similarity [12], which allows the researcher to understand the logic of errors and changes when correcting the model; -a technique for synthesizing software component models [15], which makes it possible to observe static properties [16] when building a component structure from PN-elements and PN-patterns, to display on this structure the peculiarities of the functioning of the model using dynamic labels;…”

“…To build and conduct an in-depth analysis of software models, a combined approach to simulation modeling of systems with parallelism is proposed [12], which involves building a model of a software tool based on a Petri net. The specified model has an unambiguous mathematical notation, which allows analyzing static and dynamic properties analytically.…”

“…Also, the model is built according to the principle of structural similarity, which allows for visual analysis during the construction of the model and its simulation modeling. This is important [12] both when designing software components and when assembling them into a complete software product, as well as when analyzing the implemented software.…”