2008 COMP 522A Modelling and Simulation Projects

The purpose of a project

The course has given a birds-eye view of different formalisms: Causal Block Diagrams, Petri Nets, Statecharts, DEVS, Forrester System Dynamics, ... These formalisms each have their strengths: some (such as Petri Nets) are more suited for analysis across all possible behaviours of a system, others are ideally suited for performance analysis (such as GPSS) whereas others allow for simulation and system synthesis (such as Statecharts). During the course, the similarities and differences between these formalisms have been shown.

To study truly complex systems, it is necessary to go beyond this limited set of formalisms. On the one hand, this means looking at other formalisms. On the other hand, this means combining formalisms. A continuous (e.g., CBD) and discrete-event (e.g., DEVS) formalism can for example be combined to form a "hybrid" formalism. Another example is the combination of some notion of spatial distribution with a known formalism such as state automata. One such combination leads to cellular automata.

The purpose of the projects is to let you explore these new formalisms. During the project presentations, you will be exposed to a large number of new formalisms which were not covered during the lectures.

There are two types of projects:

• building a simulator/analysis tool for a new formalism
• using an existing formalism/tool to study an interesting problem by means of modelling and simulation

The structure of a project

• All material related to your project needs to be posted on the course web-page (to provide access to future generations of COMP 522 students). You will be given an account on the MSDL web server (look on WebCT for the login name and password). Just follow the link (with your name between [ ]) next to your project. Note that you have to append /manage to the URL of your project page to be able to edit it.
• Your project website should include the proposal, the solution (code as well as discussion), references ...
  Have a look at this nice example.
• Requirements/Design/Implementation/Tests
• Possibly iterate over multiple prototypes
• All implementations in Python or a Modelling and Simulation formalism/language/tool (except when explicitly agreed otherwise)
• Milestones/deliverables:
  1. Discuss/negotiate project subject with instructor by Friday 7 November
  2. Post project proposal + present/discuss in person with instructor/TA
  3. Progress discussions (weekly)
  4. Post presentation, report and sources
  5. Final presentation (+ demo) in class (on Monday 15 December). Note that you need to attend the presentations given by your colleagues.

Examples

• Projects of previous incarnations of this course: 2006, 2005, 2004, 2003, 2002, and 2001.

• Hybrid (continuous-discrete) modelling and simulation
  • DEVS + Causal Block Diagrams (applications: bouncing ball and train model) [Mathieu Lavallée-Adam and Mathieu Rousseau].
  • Statecharts + Causal Block Diagrams for Variable Structure modelling.
• DEVS
  • Improved pythonDEVS simulator [Fei Men]
  • Distributed classic DEVS simulator with TimeWarp [Amr Al Mallah].
  • Symbolic transformation of Coupled into Atomic models (classic DEVS, using muModelica compiler) [Jesse Doherty].
  • DEVS simulator in Modelica (using "when").
  • Emulating classic DEVS in parallel DEVS.
• Causal Block Diagrams (CBD)
  • An optimizing compiler for hierarchical Causal Block Diagrams. Application: the PHYSBE benchmark model of the human circulatory system [Nurudeen Lameed].
• Cellular Automata (CA)
  • A Cellular Automata simulator for the GraBaTs AntWorld Simulation case [Jessica Li On Wing].
• Petri Nets
  • Analysis of Place/Transition nets (conservation via matrix representation).
  • A Coloured Petri Net simulator.
  • Analysis of Timed Petri Nets.
• Process Algebra
  • Type checking/inference and analysis for a small timed, process algebra-like language [Darin Morrison].
• Analysis of Statecharts + Sequence Diagrams through mapping on Place/Transition Petri Nets [Wisam Al Abed].

Using an existing simulator/analysis tool in a particular application domain

• Distributed Physics System performance analysis using DEVS [Tom Chen].
• Discrete-Event Modelling of Computer Architecture. In particular, design, from basic building blocks, the datapath for a simple 8-bit processor (specification).
  • using the visual modelling and simulation tool logisim [Farah Fariz] [Victor Cano Becerril].
  • using DEVS (for a sub-set of the full specification)
• Modelling (emergent) agent behaviour.
  • Study the behaviour of a spatially distributed system: disease spreading.
    • using collections of interacting Statecharts [Nabeelah Ali].
    • using a lattice of DEVS models [Jean-Philippe April].
  • Study flocking of boids using collections of interacting Statechart in Repast [Diana Gheorghiu].
  • Study behaviour of ants using a discrete-time, cellular agent model in Repast [Fabian Kalin].
• Use Coloured Petri Nets with time (CPN Tools) to model network routing [Amin Ranjbar].
• Use the multi-formalism modelling and simulation tool Ptolemy II to study a personalized rapid transportation system [Onur Duman].
• Model simple aircraft dynamics using CBDs. Use behaviour trace to animate behaviour in FlightGear (see also The Mathworks' example) [Elijah Johnson].
• Model and synthesize in-browser user interfaces (for Domain-Specific Modelling) with Statecharts [Raphael Mannadiar].

Original Work

You are encouraged to help each other formulate the ideas behind projects and find inspiration in the literature and on the web, but each team is required to submit their own original work. Handing in work that is not your own, original work as if it is your own is plagiarism. All re-use, collaboration, inspiration must be explicitly mentioned in the assignment.
McGill University values academic integrity. Therefore all students must understand the meaning and consequences of cheating, plagiarism and other academic offences under the code of student conduct and disciplinary procedures (see www.mcgill.ca/integrity for more information).