Yee Han Ang and Laura
Antarctica New Zealand (hereinafter referred to as ANZ) provides support for visitors going to Antarctica, either for science or non-science events. Science events deal with scientific studies of the environment in Antarctica and non-science events include art-related events, photography projects, media & invited visitor events etc. The support provided by ANZ include transportation to and from Antarctica as well as within Antarctica. ANZ also supplies food, clothing, shelter and equipment to events. New Zealand’s station down at Antarctica is called Scott Base.
Scott Base is the transition point to the field and they have staff there all year round. However, the season in which events are scheduled only runs from October to February. During this period, Scott Base staff have to accommodate preselected science and non-science events with their limited resources.
There are several problems that limit ANZ’s ability to effectively schedule events.
Our main objective was to design a knapsack/scheduling program to select events and schedule them to check for resource capacity feasibility. The main purpose of this tool is to reduce the effort in event selection. As stated in the problem situation, this is currently done manually, and hence is very-time consuming. If there are changes made in certain event requirements, the process of incorporating these changes in the selection and scheduling has to be done again by hand, making the whole task very tedious.
The program is an integer program with binary variables and is a combination of a knapsack and scheduling problem. We select some events from a pool subject to capacity constraints and then schedule the selected events. There are a few assumptions to the model.
Maximise total values of selected science and non-science values.
Note that these values are equivalent to priorities of events that will be supplied by ANZ.
We designed our model to run in LP_Solve IDE, a free solver that can be downloaded from the internet. Data entry is done through Microsoft Excel 2007 and VBA code is used to transform this data into a .mod file that can be solved in LP_Solve. The output from LP_Solve is then transferred back into Excel for easy interpretation. Part of the data entry spreadsheet is shown below.
The following diagram shows tables with data for the season as opposed to each individual event.
Output is produced in two different sheets. The diagram shown below shows part of the schedule as obtained from the LP_Solve solution. The black cells show the dates on which an event will take place. For example, event K003 will run on the 6th of October and its duration is only one day.
Another form of output shows the total resources used by all events per day and over the whole season. The pop-up labelled “Resource File Name” enables the user to enter the name of the file in which the LP_Solve output is stored at.
We would like to thank Greg Jack from Antarctica New Zealand for his
help in obtaining data and explaining the details of the records and the operations of ANZ to us. We would
also like to thank all other staff of Antarctica New Zealand who placed a lot of faith in
our project as their excitement and encouragement helped a lot in motivating
us to put in continuous effort in improving our deliverables.
Finally, we would like to thank our supervisors Dr John Giffin and Dr Nicola Petty for their support and advice.