(C) 2018 Jan Kumor, Patryk Kocielnik, Warsaw University of Technology

Type what comes below in the Jupyter Notebook.

%matplotlib inline

e = resiliences.experiment(
            vertex_count = 10^2,
            edge_density = 10^-1,
            edge_distribution = "random"
        )

e.save_plot("plot.svg")

When routine finishes, collect the plot from the execution directory.

For the current step to be complete, the commands below should:

• Display an averaged graph here in the notebook, and
• Save the graph under the name given in the working directory.

%matplotlib inline

e = resiliences.experiment(size = 100, density = 0.1, graph_type = "random", epochs = 10)
e.save_graph("graph.svg")

Currently, the notebook looks as presented in the image.

The class with experimental methods is src/resiliences.py.

The main notebook for the experiment is src/report.ipynb.

To run the notebook you need Jupyter Notebook (formerly iPython Notebook). Install with pip install jupyter.

The report (the deliverable) is in the main directory of the project in the PDF file named report.pdf.

To generate the report from files located in directory in where the report draft is located, issue the make command while in the project root directory.

Correct results in the graph below is what we are after. Then we need to aggregate similar results for varying graph sizes and connectivity factors (densities, k).

The curve should grow with the horizontal variable, but it is yet unknown, at what proportion (if at any proportion at all, vide the notion of scale-free nets).

Directory tree:

• workshop - Data needed to generate the deliverables (the project engine),
  • contract - Requirements for the project,
  • in - Source text of the report,
  • out - Destination for the generated report,
  • src - Source code of the experiment (a Python class + sauce & spinach).

The report is generated from Markdown.

This approach saves a lot of lots in terms of generation time. To withess this miracle, say cd workshop and then make.

W przypadku sprawozdań spóźnionych system automatycznie naliczy karę w wymiarze 10% punktów/dzień.

Poprawie nie podlegają sprawozdania spóźnione ani takie, w stosunku do których zachodzi podejrzenie, że zostały załadowane do systemu wyłącznie w celu uniknięcia kary za spóźnienie. W tym drugim przypadku wobec Autorów mogą zostać wyciągnięte dodatkowe konsekwencje. Tryb ewentualnej poprawy sprawozdania należy ustalać indywidualnie ze sprawdzającym.

1. NetSwan: A set of functions for studying network robustness, resilience and vulnerability
2. CyGraph: Big-Data Graph Analysis For Cybersecurity and Mission Resilience

1. Gather results for $n$ from range [10 .. 100 000] and densities of [0.001 .. 1]
2. Use multiprocessing lib to speed up computations.