Optimisation postgraduate personal statement example

Personal statement example

Every Thursday I sit down with a paper diary, twelve part-time contracts, opening hours, delivery slots and a note of who has asked for Saturday off, and I build a rota for the garden centre where I work. It takes about an hour, it is never quite right, and it was the first thing I recognised when I met the assignment problem in my second undergraduate year. Staff scheduling is not what drew me to mathematics, but it is what convinced me that optimisation is where I want to work: the problems are ordinary, the structure is deep, and the difference between a good and a bad solution is something people actually feel.

My BSc in Mathematics gave me the foundations I needed, and I chose modules deliberately once I knew where I was heading: linear algebra, analysis, numerical methods, and an operational research option covering the simplex method, duality and network flows. Duality was the idea that reframed the subject for me. Learning that the shadow prices of a linear programme carry economic meaning, and that my rota constraints each have a price attached, made me read further into convex optimisation. I worked through parts of Boyd and Vandenberghe's Convex Optimization in my final year, mostly the material on convex sets and functions and the Lagrangian; I found the KKT conditions hard going and returned to them several times, which is honest but also the reason I want formal teaching rather than continued self-study alone.

My final-year project modelled deliveries for an invented small bakery supplying six cafés. I set it up as a vehicle routing problem with time windows, then simplified to a single-vehicle formulation I could actually solve, using Python with PuLP and the CBC solver because I had no access to commercial software. That constraint turned out to be instructive. I could not brute-force my way past a poor formulation, so I spent real time on the model itself: tightening subtour elimination, scaling the distance data, and testing how sensitive the solution was to one café changing its opening time. I wrote up the failures as well as the results, including an early version that ran for forty minutes and returned a route no driver would tolerate. My supervisor's main criticism was that I had not compared my exact solution against a simple heuristic, and I have since read enough about local search and metaheuristics to see why that comparison matters.

Outside work and study, I score for a local cricket club, which has made me comfortable with recording data accurately under mild pressure and with explaining numbers to people who did not ask for them. At the garden centre I train new staff on the till system and handle stock ordering for the seed and bulb section, where I have started keeping a simple record of what sells in which week rather than reordering by instinct. It is not sophisticated forecasting, but it has reduced the amount of unsold stock we write off in autumn, and it has made me curious about how inventory decisions and scheduling decisions interact.

From an MSc I want rigorous treatment of integer and combinatorial optimisation, stochastic programming, and the algorithms behind the solvers I have only used as black boxes. I would particularly like a dissertation on scheduling or routing with uncertain demand, since that is closest to the problems I have met. In the longer term I would like to work as an analyst in logistics or public service planning, where modelling is done for people who have to live with the timetable afterwards. I am prepared for the mathematical demands, and I am used to fitting sustained study around paid work.