Research Theme 1
Understand the interdependencies within the infrastructure networks, along with carrying out economic and risk assessment with the use of AI.
Did You Know?
The power grid also has to worry about cybersecurity issues!
Main Goals
1.1
Find and learn what relies on the energy infrastructure network and what it relies on. Afterwards, use that knowledge to assess the risk and economic impact of possible failures.
1.2
Study the vulnerability of the energy networks, along with any clusters of nodes they are potentially vulnerable. Then look into consecutive failures and evaluate the risks and impacts of it all.
Gather
data to create a clustering system that is predicated on multi-layered network embedding.
Test
the impacts of crucial faults and following through with economic and risk assessments.
Learn
about vulnerabilities and consecutive failures, along with more risk and impact assessment.
Validate
and create algorithms to pinpoint certain node clusters that are vulnerable.
Current Objectives
Approach
It is very important that we are able to identify the different interdependencies, weak, strong, and hidden, among the industries and infrastructure network to scope out the possible risks and vulnerabilities they may have. The factors that we will be incorporating into this include cross-domain interactions, dynamic network load, and the socio-economic impacts. To achieve this we will be exploring multi-layered network embedding to create a clustering framework of our own.
We are going to look at our infrastructure network as two interconnected layers of networks. The first layer will make up the "Energy Network" and will consist of sublayers to represent the oil, gas, and electricity distribution networks, while the second layer will make up the "Economic Network" that include many intertwined industries.
​
The nodes in our layers will supply us with heterogeneous data with multiple attributes. Different cross-layer network dependencies, and within-layer connections will also be brought into the system. What we want for our framework is to be able to gather rich node information while also grabbing their interconnections to create multi-layered interdependent networks.
​
Being that the topic of obtaining embedding representation on multi-layered networks is not one with much research, we will be looking at this problem as one of optimization. With the use of different objective functions we can gather the rich information we need, the within-layer connections, node attributes, and cross-layer dependencies. Once completed, we can use the representation as input to other clustering algorithms to identify certain clusters within the infrastructure networks!
Did You Know?
The energy network also depends on industries like communication, transportation, and finances.
Did You Know?
CEP is a relatively new process that can match incoming data to patterns that discern for us what is happening.
Risks will be looked at on both the macro and micro levels. At the micro level we will choose rural towns and acquire energy network failure data that consists of the frequency, cause and duration as well as the demographics of the chosen towns. This added with surveys and previous year's data will help us predict risk levels later. At the macro level we will use the complex event processing (CEP) application.
When it comes to impacts, we plan on gathering data about the impacted social and economic pursuits from a failure before and after it occurred. Then, with the use of AI, we will build a model that predicts the risk and seriousness of failure instances with economic loss in mind. Once completed, this will utilized as a feature in clustering/role evolution model.
​
The clustering/role evolution model will be created after the examination of roles in different sectors with graph and feature role discovery methods. This will help us with economical dependency features for role dynamics.
