A Depot Location Problem

for Red Bus Limited


Brendon Smith and Kingsley Tang




About the Problem


This project involves an investigation into the efficiencies of potential satellite depots for Red Bus Limited. Operating a fleet of over 200 buses from its single current depot, Red Bus Limited is Christchurch’s largest urban bus company. Its increasing fleet numbers and expected future growth give rise to the need to examine both the number and location of new operational depots. This project focuses on the business problem from a greenfields perspective, as we analyse and explore the various options on location, size and configuration of depots that would most cost-effectively meet the company’s current needs and anticipated future growth, even without the constraint or advantage of the current depot.


About the Solution


“Deadheading” time, where wages of the drivers are paid but no revenue is collected, is one of the major costs faced by Red Bus. Our investigation on different depot options aims to minimise costs associated with this deadheading time.


After analysing the current operations and taking future growth into account, we have identified a number of potential satellite depot locations. We then built a model using Microsoft Excel and VBA to assign bus blocks to their nearest depot, according to the rectilinear distance approximations between bus blocks and potential depot locations. A tie-breaking process involving straight-line distance approximations was also implemented. Minor adjustments were made to ensure these estimates reflect actual distances accurately.


The model allows the user to specify minimum and maximum capacity limits for both the current and satellite depots. On the one hand, it may be uneconomical to have a depot only serving a few blocks, but on the other hand, it may not be feasible to build a depot that is too large. A heuristic has been designed to adjust the initial block allocations, so that blocks are reallocated to different depots in order to satisfy all the capacity restrictions. The block assignments and savings in costs related to deadheading time are available to the model user to aid in their decision-making process. Furthermore, the users are able to modify aspects of the model such as the different fixed and variable operations costs, as well as construction costs.


A cost-benefit analysis and an evaluation on changes in parameters and scenarios complement the recommendation to our clients.



Figure 1. The user interface of our model, DepotPro v1.1.


About the Authors


This project has been undertaken by Brendon Smith and Kingsley Tang, as part of the Management Science honours programme at the University of Canterbury in 2007. Practical projects such as this allow students to gain experience in solving ‘real-life’ business problems by working closely with an organisation in the community. Brendon and Kingsley will be moving into business consulting upon graduating from the honours programme.




Brendon and Kingsley would like to thank our clients at Red Bus Limited, for providing us with an interesting and challenging problem. They have also kindly supported us with answers to our many questions during the year, and data to apply in our model.


We would also like to thank our project supervisors, Dr Don McNickle and Dr John Giffin, who provided invaluable advice and guidance over the duration of the project.


A special thanks goes to Xijing Li, who shared the initial part of this journey with us. We wish you well in your recovery.