Features

Feature Overview of TimeNET

Overview

TimeNET is a software tool for the modeling and performability evaluation using stochastic Petri nets. Its main characteristics are the evaluation of models with nonexponentially distributed firing delays, the ability to model and evaluate complex colored stochastic Petri nets, and efficient simulation methods for models with rare events.

TimeNET has been successfully applied in several modeling and performance evaluation projects and there are several hundred installations in universities and other organizations worldwide.

Extended Deterministic and Stochastic Petri Nets - eDSPNs

The standard model class in TimeNET contains places and transitions as well as input, output, and inhibitor arcs. Transition delays may be zero (immediate), exponentially distributed, deterministic, or general (defined with a special syntax). The net class contains classic SPNs and GSPNs as subsets.

Available Features:

  • Interactive or automatic token game with break points to debug models
  • Structural analysis techniques: place and transition invariants, extended conflict sets, state space size estimation, traps, siphons
  • Numerical stationary analysis (steady-state) and transient analysis
  • Stationary simulation: including parallel replications, batch means, and RESTART splitting
  • Experiment feature: automatic solution with varying parameter settings and graphical plots
  • Graphical display of the reachability graph and token distribution probabilities
  • PNML import and export of models

Stochastic Colored Petri Nets - SCPNs

This model type combines colored Petri nets with freely definable object-like tokens with the probability and time concept generally accepted for stochastic Petri nets. The tool is thus capable of modeling complex systems in a compact way.

Implemented Capabilities:

  • Interactive or automatic token game to validate intended behavior
  • Token color definitions similar to object-oriented programming (structured types and references)
  • Stationary and Transient simulation with accuracy control
  • Rare-event simulation of SCPNs to speed up long-running reliability problems
  • Complex result measure definitions (all events or color-selective)
  • Database interface for data-intensive model settings
  • Full syntax check and display of incorrect model elements

Automata / Markov Chains

A simple net class for classic Markov chains is available in TimeNET (primarily for teaching purposes). State and state transitions with their rates or probabilities can be drawn, and performance measures computed.

Teaching & Analysis Features:

  • Models interpreted as discrete or continuous time (DTMC / CTMC)
  • CTMC stationary solution
  • DTMC stationary solution
  • Parameters that depend on other parameters

Results of transient evaluations and experiments as well as intermediate measures that can be plotted are visualized in the Result Monitor. This graphical backend receives data from the analysis modules and displays them as plot graphs, which can be widely configured by the user.

Models can be exported to SVN and PDF for printing or insertion in your publications. All models are stored in a human-readable XML format and based on TimeNET’s internal net class definitions in schema files. Moreover, the evaluation modules can be used from the command line or scripted without the graphical user interface.

The software architecture contains a Java graphical user interface and many evaluation modules written in C++ and C, which communicate via intermediate files and sockets and are compiled for Windows and Linux from the same source tree.

Sample Application Areas

  • • Communication systems
  • • Embedded systems
  • • Reliability evaluation
  • • Train control systems
  • • Manufacturing
  • • Supply chains
  • • Business processes