Operations Research Workshop at the University of Cape Town

By Klara Engelbrecht, Charlize Kriek, and Rohan Naidoo

Last month, from the 10th to the 12th of September, an Operations Research Workshop was hosted at the University of Cape Town. This event was open to academic staff, as well as Honours, Masters, and PhD students interested in applied decision-making and analytics. The program included talks, tutorials, and hands-on sessions featuring speakers from academia, government, and industry. Each day was divided into two morning sessions and one afternoon session, with tea and lunch breaks in between. These breaks provided ample opportunities for networking and discussion.

Day 1: Engaging Presentations and Insights

Wednesday morning kicked off with a presentation from Prof. Franklin Djeumou Fomeni. He shared four different operations research projects he has worked on, including air traffic management, tea blending in the UK, water supply optimisation, and smart logistics systems. These projects served as striking examples of how operations research (OR) can create real-world impact.

The second morning session was led by Prof. Sheetal Silal. She introduced the different types and uses of web-based applications. Prof. Silal explained how these applications can function as dashboards that tell a story with data, engines that allow for strategy testing, facilitators that evaluate the impacts of changes, and even debuggers that help spot errors in code visually. Following her talk, Mr. Retselisitsoe Monyake conducted a live tutorial on building applications with the Shiny package in R-Studio, inviting the audience to code along. Mr. Monyake also recommended Mastering Shiny by Hadley Wickham as a great resource for those looking to dive deeper.

In the afternoon, Dr. Lieschen Venter delivered an insightful talk about what a career as an OR graduate might look like in various sectors: private, public, or academia. She laid out the pros and cons of each path. The private sector offers the satisfaction of seeing your work implemented quickly but comes with inherent risks. Academia provides a stable path and intellectual freedom but carries constant publication pressure. The public sector allows for social impact, though political risks and ethical responsibilities can complicate the work.

The day concluded with an interactive session led by Dr. Robert Bennetto. He revisited a previous project where a South African coal transport company needed an automated scheduling tool for its trains. Understanding the problem took months, and the project raised numerous data and ethical challenges. The audience debated which methods they would have tried—simulation, MIP, heuristics, or metaheuristics—before hearing how the project actually unfolded. This lively discussion wrapped up the day and illustrated how technical modeling and human decision-making often collide in practice.

Day 2: Deep Dives into Industrial Applications

On Thursday, Dr. Robert Bennetto began the day with an engaging session on Industrial Scheduling. He focused on exact methods for solving the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW). Dr. Bennetto elaborated on various constraints and cutting methods applicable to this problem. After discussing the theoretical formulation, he shared practical applications, including a waste collection scenario and a mixed in-source–outsource scheduling case.

In the afternoon, Prof. Renette Blignaut presented a thought-provoking session on the Use of Artificial Intelligence in Operations Research (OR). She highlighted the contrast between traditional OR, which is typically structured and deterministic, and AI-enhanced OR, which leverages adaptive and dynamic models. A major discussion point centered on the ethics of AI use, particularly the importance of maintaining human judgment. Blignaut emphasised that practitioners should not rely solely on AI outputs but also apply intuition and critical thinking, questioning whether the model or its results truly make sense.

Dr. Jonas Stray then presented on Analytics in Retail. He began by outlining the structure of analytics teams within the retail sector and how they collaborate with business executives. Dr. Stray discussed a range of retail-focused analytics projects, including demand planning, price optimisation, distribution allocation, target arrival optimisation, markdown optimisation, and safety stock management. He concluded by showcasing the tangible impact of these projects, with some initiatives achieving profit improvements of up to 12%.

The day ended with an introduction to systems thinking and dynamics by Dr. Lieschen Venter.

Day 3: Implementing Systems Dynamics

We began the final day right where we left off on Thursday, focusing on implementing what we learned about systems dynamics. Dr. Venter guided us through the process of representing complex behaviours and the paradigm shift required to think in this way. With the knowledge we had acquired, we built a small model in AnyLogic, a powerful tool for simulation. This exercise allowed us to identify areas for improvement in the system and see how distinct parts of a system can affect the outcome.

After a tea break, Dr. Isabel Meyer presented her work at the CSIR. She delved into the different strategies employed in the institute’s decision-support applications. Her talk shed light on the role of the CSIR as a trusted intermediary that collaborates with various bodies to contribute to existing knowledge. A great emphasis was placed on the importance of soundly structured problems and the understanding that can be gained from well-constructed models.

Dr. Meyer’s talk served as a suitable introduction to Prof. Leanne Scott’s session on overviewing problem structure methods. The presentation clarified the purposes of soft and hard OR and displayed the usefulness of techniques like MCDA through interesting case studies. Once again, attention was drawn to the importance of problem structuring, along with the myriad nuances that accompany it. A key takeaway was the necessity of maintaining stakeholder involvement within the decision process, especially in solving problems where these stakeholders are directly affected.

After a satisfying lunch, we entered our final session of the workshop. Prof. Fomeni, who had presented on the first day, briefly introduced us to the quadratic knapsack problem. What followed was a fun competition to see who could develop the best sorting criteria algorithm in the time we had left, with the winners receiving a cash prize.

Conclusion: A Valuable Experience

In conclusion, the workshop was a wonderful event for many soon-to-be OR graduates. The opportunity to observe practical implementations and engage with seasoned professionals was of immeasurable value. Heartfelt appreciation is due to the speakers and to UCT for hosting us!

The insights gained from this workshop will undoubtedly shape our future careers in operations research. The blend of theory and practical application provided a well-rounded experience that we will carry forward into our professional lives.