May Hicks Award


The OR Society funds its annual awards for student projects from a generous bequest from the estate of Mrs May Hicks, wife of Donald Hicks OBE, a major contributor to operational research and the first treasurer of The OR Society. Projects entered are OR projects carried out for a client organisation rather than within the university.

Entries for any year should be submitted electronically to [email protected] to arrive no later than 30 April.

Citation for the May Hicks Awards 2025

The OR Society is delighted to announce the winners of the 2025 May Hicks prize for Best Post-graduate Project.

The winner, Lei Zhou (University of Southampton), received £1000. The runners up were Apichaya Chanmathikornkul (Lancaster University) and Mouri Hoque Nadia (University of Southampton) who received £250 each.

Thanks go to the Universities who submitted their student projects. It was pleasing to read about such a range of business problems and how well they are being addressed by OR techniques.

Winner:

Lei Zhou (University of Southampton)

Project ‘Optimizing Production Efficiency and Reducing Costs in the Frozen Vegetable Food Processing Industry’

This project focused on developing a data-driven approach to optimise production planning in the frozen vegetable processing industry, addressing the operational challenges faced by Ittella Italy Srl.

Lei developed a mixed-integer linear programming (MILP) model to support lot-sizing and scheduling decisions, tackling the core difficulties of seasonality, raw material perishability, and complex production constraints. Her work enabled Ittella to transition from intuition-based planning to a more systematic, cost-effective approach. The project had two main components:

  • A MILP-based optimisation model to determine optimal production schedules and quantities, minimising total costs while meeting product demand. The model took into account Ittella’s production process, capturing key operational constraints such as harvest windows, setup times, shelf-life limits, and machine maintenance requirements.
  • Implementation and testing of the model using real-world scenarios derived from Ittella’s operational data, demonstrating significant cost savings, reduced processing time, and improved resource utilisation.

Frozen vegetable production is uniquely constrained by biological factors, with narrow harvest windows and highly perishable inputs. Lei’s model provided Ittella with clear, actionable production plans that reduce waste, avoid costly downtime, and improve the use of labour and equipment across the season.

This work directly supports Ittella Italy’s ambition to modernise its operations and increase resilience in a competitive, seasonal market. By using optimisation to align production with natural harvest cycles, the company can meet customer demand more reliably and efficiently.

Lei’s research demonstrates the value of applying advanced operations analytics in the food manufacturing sector and paves the way for broader adoption of evidence-based planning methods in industries with similar seasonal and perishable constraints.

Runners-up

Apichaya Chanmathikornkul (Kat) (Lancaster University)

Project ‘Demand Forecasting for Metroplan Limited’ 

Kat’s dissertation focused on improving demand forecasting at Metroplan Limited, where ineffective forecasting can result in stock imbalances. These inaccuracies lead to costly overstock or understock situations, which in turn affect both operational efficiency and customer satisfaction.

The dissertation presents a systematic approach that applies time series forecasting methods tailored to different types of product demand. Demand is first categorised using the Syntetos-Boylan-Croston (SBC) method, allowing for the selection of forecasting models best suited to each pattern—whether smooth, erratic, lumpy, or intermittent. Six forecasting techniques were explored, including Exponential Smoothing (ETS), Croston’s Method, and the Teunter-Syntetos-Babai (TSB) method.

The results show that matching forecasting methods to demand type improves accuracy by up to 25.69% compared to Metroplan’s current approach. The dissertation also explores the trade-off between applying one forecasting method across a category versus selecting the best method for each product individually, highlighting the benefits of a more tailored approach.

The project culminates in the development of an R-Shiny application that automates the entire forecasting process. This tool is designed for non-technical users and enables quick, accurate forecasts through a user-friendly interface. The application handles data preparation, demand classification, and forecast generation, with built-in visualisations and prediction intervals to support informed decision-making.

The dissertation reports strong results and demonstrates the value of automated, data-driven forecasting. The tool provides a practical solution for companies like Metroplan seeking to improve inventory planning and reduce stock-related costs. This development is poised to enhance business efficiency, support customer satisfaction, and contribute to more strategic supply chain decisions.

 


Mouri Hoque Nadia (University of Southampton)

Project ‘Quality Data-Driven Sustainability: Optimising Spare Parts Manufacturing and Inventory Management for a zero-Waste Future’

Nadia’s dissertation focused on automating the classification of Environmental Product Declarations (EPDs) in the bearing industry, where current processes are manual, time-consuming, and inconsistent. As green-taxonomy rules tighten, companies like KOIOS Master Data face increasing pressure to process large volumes of complex EPDs quickly and reliably.

The dissertation presents a machine learning approach that reframes EPD assessment as a clustering problem. Two unsupervised learning methods (hierarchical and spectral clustering) were benchmarked on seven real-world datasets, with spectral clustering revealing a more accurate four-tier environmental impact structure. 

The solution significantly reduces processing time, enabling EPDs to be classified in under five seconds compared to two hours manually. It also improves consistency, as the clustering is driven by deterministic algorithms rather than subjective human judgment.

The project delivers a practical tool that KOIOS now pilots as “EPD-Rank,” supporting its clients—particularly SMEs—with affordable, auditable sustainability scoring. This work not only streamlines a critical supply chain workflow but also demonstrates how advanced clustering techniques can be applied to environmental data in a business context.

The dissertation reports promising results and contributes a novel application of spectral graph methods to life-cycle assessment. It provides a strong foundation for further research while offering immediate commercial impact. This development is poised to enhance sustainability reporting, support regulatory compliance, and equip SMEs to compete more confidently in an environmentally focused marketplace.