Engineering Science

Project 

Groundfish surveys collect data on the abundance, distribution, and biological characteristics of fish populations. To do this vessels must visit many stations in the ocean and collect samples. The design of the routes that the vessels take between stations and ports, to unload the caught fish, is a form of capacitated vehicle routing problem.

The stations are two distinct points that must be traveled between, some in any direction and others in a particular direction (due to the weather or ocean floor conditions). In addition the amount of fish that will be caught at each station is not known beforehand. These factors make the problem difficult to model and solve.

The role

There are several potential avenues for exploration in this project:

• A simulation model that captures the complexity and uncertainty of the problem, so that solutions can be evaluated in a realistic setting, including replanning decisions
• Developing an exact method for creating initial plans based on column generation and branch-and-price
• Improving current methods for predicting the catch at stations

Ideal student

An understanding of linear and integer programming, and Python development proficiency would be beneficial for this project.

Project 

In terms of Discrete Event Simulation (DES) a conceptual model is a formal representation of the system being modelled, and acts as an intermediary step between the real world situation and the computer model. Hierarchical Control Conceptual Modelling (HCCM) is a framework that defines a strict standard that conceptual models should adhere to, which improves the ability to communicate conceptual models unambiguously.

Currently both developing a conceptual model that adheres to the HCCM standard, and implementing it in a DES software, are difficult, error-prone, and time-consuming tasks. For the conceptual model, all of the components of the system, and the relationships between them need to be specified precisely. For the simulation model, you must make sure that the conceptual model is represented accurately.

The role

This project would investigate the use of a software tool (or tools) to automate parts of conceptual model development and simulation implementation. The aim is for users to be able to construct the activity diagrams of the entities in the system. Then the tool creates all of the necessary structural and relational information from the activity diagram. Users would then only need to add any additional details or make minor modifications.

The tool should also be able to convert the conceptual model to various formats, for example the basis of a model file for a simulation software, or formats designed for comprehension such as pdf or HTML.

Ideal student

Prior experience with conceptual modelling for DES, and the HCCM standard in particular would be beneficial.

Project 

Non-motorised modes of transport such as public transport, walking and cycling are increasingly important components of a more sustainable transport system. Higher use of non-motorised transport can help improve traffic congestion, and it increases resilience to rising oil prices.

A model to determine cyclist route choice has been developed based on the assumption that cyclists seek a short but 'high-quality' path, where 'quality' refers to a multitude of attributes such as quality of infrastructure, traffic volumes, safety, etc.

We will develop a routing approach to identify good cyclist routes trading off directness of route and 'quality' by modelling cyclist route choice as a bi-objective shortest path problem where cyclists are assumed to choose paths that represent an optimal trade-off between distance and quality.

The role

Based on this, one can model expected cyclist flow along various roads or cycle paths in a road network. We will use the above cyclist route choice model to understand where the current cycling network has gaps or bottlenecks, and where to prioritise network investment.

To achieve this, we will draw on various metrics used in social network analysis to investigate whether those network centrality measures (and their bi-objective versions) can provide meaningful insight into what the most critical cycling network infrastructure may be.

Aims

• Develop a cycling network for a large city such as Auckland, or perhaps two different cities for comparison. This will likely be done using open-source road network data from OpenStreetMap or Overture Maps.
• Develop bi-objective shortest path algorithm to derive paths that provide good trade-off between distance and quality.
• Implement centrality measures based on bi-objective shortest paths and use them to analyse cycling networks

Ideal student

Note that this project is suitable for Engineering Science Operations Research (OR) students who are confident in programming, which is required in all aspects of the project. It may also be suitable for students in other Engineering disciplines who are confident programmers and familiar with algorithms such as shortest path search or A* search. Existing C code for solving bi-objective shortest paths is available. However, it may be preferable to develop new code written in Python or Julia.

Project 

Before we are born, the cardiovascular system develops ready for later life. But, the system functions differently in the fetus, due to our supply of oxygen to grow coming from our placenta, and not our lungs.

The role

This project aims to understand how cardiovascular development differs in pregnancies where the fetus does not grow as it should. The project will use state-of-the art microCT data to develop finite element models representing the fetal cardiovascular system's anatomy (with a focus on heart and placenta at the end of pregnancy in rodents) and simulate tissue function using open-source modelling software.

Ideal student

The project will suit students who have an interest in mathematical and computational modelling as well as medical imaging and image analysis.

Requirement

Knowledge of programming (e.g. python/C) is essential.

Project 

There are several mobile and web applications available in pregnancy, where expectant parents can see expected growth and development of their baby. But, these applications often gloss over pregnancy complications, which although rare, require significant monitoring of pregnancy. Doctors and midwives struggle to explain to patients the physiology that is associated with pregnancy complications, which relates to the reason that a person may need to attend additional appointments of tests.

The role

This project aims to build upon a web-based application that uses "virtual pregnancy" models – simulations of physiology in a real pregnancy – to educate about pregnancy health and monitoring.

Requirements

The project involves web development using an existing technology stack used at ABI (primarily Vue.js, Nuxt.js, Python, Github) so requires strong knowledge of one or more of these technologies.