Academic Interview - ME205: Strategic Decision Making: An Introduction to Operations Research Methods

We sat down with Dr Ahmad Abdi and Dr Katerina Papadaki, who teach ME205: Strategic Decision Making: An Introduction to Operations Research Methods, to learn more about the impact and relevancy of Operations Research, and the ways in which students can use it for future strategic decision making.

 

What are the key characteristics of Operations Research, and why is it relevant to today’s world?

Most business decisions today aren’t made capriciously by managers but driven by processes and data. Operations Research (OR) addresses quantifiable problems, often involving large-scale data, to make better decisions and improve the effectiveness of operations and management. It is a multidisciplinary approach which relies on mathematical models, optimisation techniques, and statistical analysis.

Operations Research is an essential toolkit in today’s data-driven, fast-paced world. Its ability to transform complex problems into actionable solutions makes it a cornerstone for strategic decision making in governments, healthcare, and industries such as finance, transportation, and supply chain management.

Aviation is a notable industry where OR has made a significant impact. By leveraging OR techniques, airlines have optimised various operational and decision-making aspects of the industry such as flight scheduling, crew assignment, maintenance planning, and revenue management, thus leading to annual savings of hundreds of millions of dollars.

What is the most fascinating thing students will learn in the classroom?

You have all solved a Sudoku puzzle before, but have you done this by a computer code before?

In this course, you will learn how mathematical modelling and optimisation techniques can assist in making many interrelated decisions in the presence of scarce resources and hard planning constraints.

Loosely speaking, the Sudoku puzzle is an instance of such a decision-making problem where the blank cells of a partially completed 9-by-9 grid must be filled with numbers from 1 to 9, in a way so that every row, every column, and certain 3-by-3 sub-grids receive exactly one of each number. We will see how to model Sudoku efficiently and then pass the model to a special-purpose software to solve the puzzle in less than a second!

Could you please describe the practical components of the course and how students will engage in hands-on learning in the classroom?

Students will learn when and how to model a decision-making problem by linear, integer, or dynamic programming, or as a network optimisation problem. They will see concrete applications to healthcare, finance, economics, transportation, and supply chain management. Students will also see model instances implemented on a special-purpose software used to solve the problem.

They will think about, analyse and model problems where randomness and uncertainty are intrinsic, such as the value of a stock in a financial market, monthly revenues of a company in a competitive market, or the waiting time of a customer in a physical or online queue. We will also look at concrete applications to search engines, and various real-world scenarios where queues occur.

During the lectures, students can bring in their laptop if they choose to and implement some of the models alongside the lecturer. Their knowledge will be tested via in-lecture quizzes. During the classes, students will have the opportunity to solve 2-3 problems, some individually and some as a group, and receive feedback from the class teacher.

How is the field of Operations Research evolving, and what are some emerging trends or innovations in this area?

Some emerging trends in the field of Operations Research (OR) include incorporating uncertainty in optimisation theory. Optimisation theory classically deals with deterministic problems, whereas chance-constrained optimisation tries to solve a problem under uncertainty and tries to offer guarantees of a certain performance despite the uncertainty.

Another emerging trend in OR is data-driven optimisation, which integrates data into the optimisation process to improve decision-making. Data-driven optimisation is widely used in various fields, including engineering design, finance, healthcare, and supply chain management.

Finally, machine learning and artificial intelligence techniques are being integrated into OR. The Travelling Salesman Problem (TSP) is a classical OR problem where a salesman needs to tour n cities, visiting each city exactly once, in a way that minimises total distance travelled. The TSP and similar problems such as the Vehicle Routing Problem have been tackled using Transformers with promising early results. A famous success story is the DeepMind AlphaGo algorithm that uses deep reinforcement learning to play Go, a strategy board game. AlphaGo has defeated multiple world Go champions! 

What sort of scenarios could students apply OR methods to in their studies or careers?

As an example, in this course, we cover critical path analysis in project management, which is one of the most useful techniques as it applies in every industry. This is when we need to schedule activities to complete a project, and the start time of some activities depend on the completion time of others. The goal is to find an optimal schedule that minimises the project completion time. We identify the activities whose completion time is critical to the project. This method can be adapted to incorporate uncertainty in the activity times.

In general, this course can help students solve problems where a lot of interrelated decisions need to be made in an optimal way; these are decisions so complex that no decision maker can make them without the help of a mathematical model and the methodology to solve it. Problems of this type can arise in the students’ careers depending on the sector they work in. Some examples are portfolio optimisation in finance, planning, production, network optimisation, supply chain and revenue management, and job scheduling.

Many real-world problems involve uncertainty and, in this course, we study stochastic processes, such as Markov Chains (MCs), that can help us model uncertainty. MCs can be used in Finance, such as stock price movement, credit risk, and market behaviour, but can also be used in healthcare, marketing, telecommunications, web search and website ranking.

Further, this course can prepare students for a career or degree in business, but it can also serve as the quantitative course option for their current degree.