La Cueva (56)/Interim Report

Interim Report
http://mode.lanl.k12.nm.us/get_interim1112.php?team_id=56

Problem Definition
Complex networks such as those in social, biological, and computing fields, are used to express connections between different nodes through links. Communities within a network can signify the connectivity between distinct groups of nodes. For example, social networks often include community groups based on common locations, relations, and interests. Citation networks have communities based on research topics. Being able to identify these communities in a network can provide insight into how network function and topology affect each other.

In this project, the goal is to compare the efficiencies of different optimization techniques on partitioning a graph to achieve high modularity.

Problem Solution
We plan to develop a model of the network which we will partition. Finding the partition of a network that maximizes modularity will be key in structuring the given network. Multiple optimization techniques will be implemented to partition the network, including the Greedy Algorithm, Ant Colony Optimization, and the Girvan-Newman Algorithm.

In order to partition the network, it will be modeled by assigning locations to the nodes based on their relationships. Groups will then form in certain areas of the model, indicating modularity.

Progress to Date
So far, a program has been written to simulate and visualize the network. A square matrix of dimension size corresponding to the number of nodes in the array is used. Each location in the array represents a single direction of a connection. Binary toggles indicate where these connections exist. The nodes are then assigned coordinate locations which will be used in partitioning algorithms.

Expected Results
We expect to find the partitioning of the modelled network with maximized modularity and detect the communities within the network. We will also determine which of the optimization techniques implemented is the most efficient at creating quality partitioning. The network’s communities being formed by the optimization techniques are displayed in a visualization in order to give more insight on the process of the formations and detections of communities.

Interim Comments
David Rogers Lead for Scalable Data Analysis and Visualization Group at Sandia National Labs

Hi. Good progress so far - can you post some images of the network visualization, with a short caption explaining them?

From your report, it sounds like you're making good progress across the breadth of the project. Also, it's great to see the works cited - that's an important element a lot of teams neglect to include.

On thing to work on - it's important for a judge to understand exactly what you mean by 'modular' and 'maximized modularity', so be sure to explain that clearly in your presentations. That's the only thing I see at the moment that isn't explicit in the report, and could significantly impact people's understanding of your results.

Good luck!