Otto-von-Guericke-Universität Magdeburg

 
 
 
 
 
 
 
 

Oral Exam Questions

This page contains a list of question related to the course "Advanced Discrete Modelling". These questions are intended to help students check that they have understood the material in the theoretical and practical classes.

Furthermore, the exam will also be based on these questions, although this does not imply that the exam will be exclusively composed of the questions listed below. In addition, you will be expected to bring the results of your homework assignments to the exam (and be able to explain them!)

Discrete-Time Markov Chains (DTMCs)

  • What is a DTMC?
  • Give an example of a DTMC
  • How are DTMCs described mathematically?
  • Draw a DTMC that describes the amount of money that a roulette player has assuming that he always bets one dollar on even.
  • How are DTMCs simulated?
  • What is a steady-state solution?
  • What is an absorbing state?
  • Explain Google's pagerank computation

Continuous-Time Markov Chains (CTMCs)

  • What is a CTMC?
  • Give an example of a CTMC
  • How are CTMCs described mathematically?
  • Give an intuitive explanation of the CTMC equations
  • Explain the "dual" (i.e. discrete-time and continuous-time) views of a CTMC
  • What is the connection between the exponential distribution and CTMCs?
  • Explain why the exponential distribution is "memoryless"

GSPNs and CTMCs

  • What is a GSPN?
  • What is a reachability graph?
  • What is the state space of a GSPN?
  • What are vanishing and tangible markings?
  • What is the connection between GSPNs and CTMCs?
  • What are the advantages and disadvantages of discrete event simulation of a GSPN compared to setting up and solving its CTMC?

Proxels

  • What is a proxel?
  • What information does the proxel carry?
  • How are proxels used to analyse an SPN?
  • What are the main properties of the proxel-based simulation?
  • For which models is the proxel-based simulation especially suitable?

Hidden Markov Models

  • What are the elements of a Hidden Markov Model?
  • What are the three basic problems that one can solve using a HMM?
  • What is a typical example of a HMM?
  • What Question does the Evaluation (Decoding, Training) problem solve? How can it be solved?
  • What are practical applications of Evaluation and Training?

Hidden non-Markovian Models

  • What is the motivation behind combining SPNs and HMMs? Practical example?
  • How would you extend the definition of a Proxel to do path & output sequence analysis?

Solving HnMM

  • How could you use the Proxel-based simulation algorithm to analyze (evaluation and decoding) Hidden non-Markovian Models?
  • How can DPH be used to train Hidden non-Markovian Models? (Idea)
Letzte Änderung: 10.04.2012 - Ansprechpartner: Webmaster