As my first term at STOR-i draws to a close I take a look back at the past ten weeks. It's been a challenging but incredibly enjoyable period, in which I have learned a huge amount. From Royal Mail problem solving days to making my own website; I've picked out a couple of the experiences which stick out in my mind.
There were two in total this term; one with Rezatec, a geospatial analysis firm based in Oxford, and one with Royal Mail. With Rezatec we were looking at how data they had from satellite images could be used to classify species of trees in mixed forests, whilst with Royal Mail we were exploring how to construct an effective delivery network for their growing parcel delivery service. Both were challenging but enjoyable.
I particularly enjoyed the Royal Mail problem solving day, in which a group of PhD students came up with a brilliant proposal. They considered optimising the use of current resources, in particular, which of the current Mail Centres to expand in order to accommodate for this new service. Part of their idea was to essentially optimise backwards, from the ideal future set up, to determine the best way to get there.
Though it sounds pretty simple, it turns out to be quite tricky, particularly when the number of items gets very large. However, after plenty time spent debugging, I managed to find solutions by coding up the Horowitz-Sahni Algorithm in R, which was very satisfying.
For now, though, I'm going to enjoy the festive period.
Problem solving days
This term I had my first experience of an industrial problem solving day. These a quite regular events on the STOR-i calendar, in which we welcome a representative from a company to the centre for the day. The representative presents us with a current problem they are trying to tackle, we then get split into groups and spend the day coming up with some ideas of approaches and solutions.There were two in total this term; one with Rezatec, a geospatial analysis firm based in Oxford, and one with Royal Mail. With Rezatec we were looking at how data they had from satellite images could be used to classify species of trees in mixed forests, whilst with Royal Mail we were exploring how to construct an effective delivery network for their growing parcel delivery service. Both were challenging but enjoyable.
I particularly enjoyed the Royal Mail problem solving day, in which a group of PhD students came up with a brilliant proposal. They considered optimising the use of current resources, in particular, which of the current Mail Centres to expand in order to accommodate for this new service. Part of their idea was to essentially optimise backwards, from the ideal future set up, to determine the best way to get there.
Packing my bags
Another part of this term I really enjoyed was the coursework I did for a module in optimisation. For this we were required to code up a branch-and-bound algorithm to solve some 0-1 Knapsack Problems. Its a very simple problem. Suppose you have a selection of items, each with given weight and profit, and a knapsack, with a given maximum capacity, then, what items should you place in the knapsack in order to maximise your profit?Though it sounds pretty simple, it turns out to be quite tricky, particularly when the number of items gets very large. However, after plenty time spent debugging, I managed to find solutions by coding up the Horowitz-Sahni Algorithm in R, which was very satisfying.
The term ahead
Next term will bring with it a new set of challenges. I will be learning how to code in C, I will be attending masterclasses in topics such as Simulation and Bayesian Decision Theory, I will also be working on two large reports and there will be further problem solving days. I will also be attending lectures on a wide range of topics in Statistics and Operational Research, after which I should begin to get a clearer picture of what areas I find interesting, guiding my choice of PhD project this summer.For now, though, I'm going to enjoy the festive period.
Comments
Post a Comment