The Essential Foundations of Operations Research
We interviewed Victor Pimenta, the current Head of Applied Intelligence at Mercado Livre, to better understand which topics in Operations Research are essential for any professional looking to stand out in the field. Below, we’ve summarized his key recommendations, ranging from theoretical foundations to advanced techniques.
1. Foundation in Mathematical Formulations
According to Victor Pimenta, the first step for anyone aspiring to work in Operations Research is to develop the ability to read and interpret mathematical formulations. This understanding aids in grasping the roles of variables, constraints, and objective functions, making it easier to analyze and solve problems efficiently.
2. Mastery of Classic Problems
In addition to reading models, Pimenta considers knowledge of classic problems in the field to be fundamental. Key examples include:
Facility Location Problem: Strategic selection of locations for facilities.
Flow Problems: Particularly flow constraints, relevant in logistics and transportation systems.
Travelling Salesman Problem (TSP): Finding the shortest route that visits all necessary points, crucial for delivery optimization.
Job Shop Scheduling: Organizing and sequencing tasks in production lines, essential for the manufacturing sector.
3. Study Reference
To deepen this knowledge, one of Pimenta’s primary recommendations is the book by Hillier and Lieberman, Introduction to Operations Research. This work provides a solid and comprehensive foundation, ideal for those wishing to go beyond the basics and explore more complex applications.
4. Decomposition Techniques
When dealing with more elaborate scenarios, it’s important to understand concepts like Lagrangian Relaxation, Dantzig-Wolfe Decomposition, and Benders Decomposition. These methods allow for large problems to be divided into smaller, more manageable parts, facilitating both analysis and the search for viable solutions.
5. Heuristics and Metaheuristics
Finally, when the size or complexity of a problem prevents an exact solution from being found in a reasonable timeframe, Pimenta suggests becoming familiar with various heuristics and metaheuristics, such as:
GRASP
Genetic Algorithm
Tabu Search
Simulated Annealing
Ant Colony Optimization
Harmony Search
Particle Swarm Optimization
Firefly Algorithm
While these approaches often do not guarantee an optimal solution, they provide near-optimal results within a practical timeframe, making them especially valuable in dynamic corporate environments.
By mastering these elements: mathematical formulations, familiarity with classic problems, understanding decomposition techniques, and knowledge of heuristics and metaheuristics. An Operations Research professional will be better equipped to tackle real-world challenges, propose strategic improvements, and effectively contribute to the success of projects and organizations.
