Apr 26, 2024
Discover How AI Can Solve Even the Most Complex of Your Problems
Digital transformation, Artificial intelligence
Throughout the day, you often find yourself optimizing your actions: you may take all your grocery bags at once to make sure you only make one trip, ask Google Maps for the shortest route to your happy hour destination, or find the perfect order to make your kids' evening routine go quickly and smoothly.
Have you ever encountered a situation where making the right decision seems impossible due to the complexity of the task, the numerous variables involved, and the countless possible solutions? That's precisely what operational research (OR) is for: to help you find the optimal solution. We explain.
In Practical Terms, What Is OR?
Operational research is a branch of artificial intelligence. It aims to provide a rapid decision-support tool by determining the best solution for a specific purpose. For instance, if your goal is to reduce the time spent washing dishes, the solution will be different from the one you will get if you want to conserve water. The operational research model can analyze numerous variables and establish how they relate to each other to optimize the outcomes based on your requirements.
How Does All This Work?
Before diving into the subject, it's important to note that humans are already great at optimization. So, when creating an OR model, we take cues from the way humans make decisions. We ask questions like "What tasks need to be optimized?" and "What is the working environment like?" to identify the overall decisions that need to be made, as well as their type (e.g., numerical or yes/no answers). We then translate this into a mathematical model, which we subject to specific constraints. These constraints are variables particular to the company that will use the solution, such as the number of employees, machines, or available materials. Then, we let our AI model solve the puzzle for us, and we get the optimal solution!
And What’s It Good For, Exactly?
You've probably already used Google Maps several times to reach the same destination. You may have noticed that the route suggested is sometimes different. Indeed, while your decision ultimately remains the same, i.e. to get from point A to point B as efficiently as possible, your environment is constantly changing. Traffic jams, roadworks, accidents... Every time you ask Google Maps for an itinerary, it has to re-optimize your route, considering many variables. But that's not all! The app can also optimize your route based on specific constraints you set, such as avoiding highways, bypassing cities, or choosing the fastest route.
Well, the RO is a bit like your GPS, but in a much more elaborate version. Moreover, its use is by no means limited to transportation! It can be employed in various fields, such as production planning, predictive maintenance, and material optimization.
To illustrate this, at Vooban, we worked on a project with a school that had to manage the schedules of around 6,000 students, teachers, and lecturers. The school needed to plan the schedules of teachers and lecturers, taking into account their various constraints, and assign them to different classes while considering their limited number as well as the equipment required for the type of course given. In this case, the number of variables to be considered was so high that it was nearly impossible for a human to solve this problem in the most efficient way. Vooban, therefore, introduced an operational research model that solves this enormous puzzle quickly and optimally!
OR requires complex calculations that increase exponentially with the number of parameters involved. While there are perfectly linear problems for which the optimal solution can be found quickly, for example, when searching for quantities, other problems, such as yes/no decisions, are more difficult for solvers and require a great deal of power and energy from the machine.
Does It Help if I Buy a Better Computer?
Yes, in theory, having more computing power can improve the chances of getting good results. However, it is more beneficial to focus on improving the algorithm, as this can generate better results regardless of the computing power available.
At Vooban, we believe in the power of diverse perspectives. Instead of solely relying on increasing computing power, we strive to include as many viewpoints as possible about the company for which we're developing the AI model. Our expertise lies here: our experts invest time to understand the customer's context and capture their reality on the work floor. This rich tapestry of insights will have the most significant impact on the optimization model.
Conclusion: OR Is the Future!
After learning the fundamentals of operational management, you're probably wondering: Does my company need it? Well, we all make decisions, and we're all subject to changing environments. So, the answer is yes! As part of running a business, decisions must be made, and with growth comes more complex choices. Every decision has a cascading effect, and it's hard to make significant gains when processes become laborious.
To explain this situation, we often use the example of a company that manufactures custom kitchen cabinets. You can imagine that the company must plan its production every day based on the received orders and associated constraints. Making a wrong decision can block several production slots and lead to a loss of potential. To simplify, automate and optimize daily production planning, this company implemented operational research and reduced its planning time from 50 to 10 hours a week!
OR comes to the rescue and helps you make the best possible decisions. Optimization is a well-honed science that's not new—it’s been around for decades! Operational research just simplifies the process of optimizing. We must learn to utilize this technology to create value as soon as possible.
And Vooban is here to guide you through it.
P.S. If you'd like to explore the subject further, we recommend you listen to the " De Zéro à RO avec Thomas Chabot " episode of the Vooban Hive podcast!
