My PhD is working with Rolls Royce to develop a Symbiotic Simulation system to help in airline disruption recovery.
Disruption due to mechanical problems, crew sickness and weather can cause severe problems to airline businesses with knock-on
effects being very costly.
When reacting to this disruption, it is important to revise the flight schedule over the immediate future to help minimise
the cost. The inherent stochasticity and complexity of the environment means that simulation modelling is a natural approach.
Symbiotic simulation refers to systems in which the simulation interacts with the system it is modelling. The simulation receives new information from measurements of the system to improve the accuarcy of the modelling, which is particularly important if the environment changes with time. The outputs of the simulation can then be used to optimise the system.
My PhD research focused on developing a simulation optimisation approach for reallocating aircraft and retiming or cancelling flights. To do this, we used a multi-fidelity modelling approach, combining the simulation with a deterministic model to guide the search.
STOR-i Internship Project
Over the summer of 2014, I was part of the STOR-i internship. My project was to look at how outliers are formed in an aggregated
The data-set I was considering was the number of faults in a network per week over a couple of years. The series was the
sum of many series, each with their own attributes such as location, job type customer type etc. The aim was to find which of the
attributes were most influential in causing an outlier. At the end of the process, I did a presentation and made a poster
summarising my project. These are both below.