Edwards hosts Mitacs international students studying AI-driven pricing to tackle global food waste
Food waste is a pressing issue worldwide. Every year, over 46% of food in Canada goes to waste.1 Globally, around one-third of all food produced is never consumed.2 In addition to lost revenues, food waste also contributes to environmental concerns. Wasting food means wasting the resources required to grow food, and food waste contributes to increased greenhouse gas emissions. But what if artificial intelligence (AI) could help solve this problem?
That question brought Paola Perez Tostado, a data science student from Mexico, and Sol Baik, a logistics student from South Korea, to the Edwards School of Business at the University of Saskatchewan. Through the Mitacs Globalink Research Internship (GRI) program, they joined Dr. Mohammadreza Nematollahi (PhD), Assistant Professor of Management Science, to explore how AI could reduce food waste and improve sustainability by offering better pricing and promotion strategies for food retailers.
The research project and skills acquired
Nematollahi, an expert in sustainable supply chain research, wants to address food waste by exploring strategies to reduce food loss and waste at the retail level.
"When we talk about food loss and waste, there are many strategies across the food supply chain, from farms and processing to retail and consumers, and ultimately back to nature as part of a circular system," said Nematollahi. "Along the way, AI can play an important role in streamlining processes and supporting better decision-making. One of the most proactive approaches is reducing food loss at the retail stage, but this is not easy. Retailers must manage inventory while dealing with perishable products that have a limited shelf life, and once products go bad, customers are no longer willing to buy them. The goal of this project was to help retailers make better decisions to reduce food loss. One practical option is using promotions or discounts before products expire. To support this, we use a reinforcement learning model as part of our AI approach to enable smarter, data-driven decisions in real time."
In short, reinforcement learning is a machine learning method where a computer model learns by trial and error. Instead of being told exactly what to do, the system figures out what works best by experimenting with different options.
"The project focused on discount strategies for perishable goods," said Sol Baik, a visiting student from Korea Aerospace University. "Our goal was to figure out the best time to apply a discount, and how much to apply in a single period, so that businesses can both maximize revenue and reduce waste. We combined a mathematical model with reinforcement learning and used Python."
"We explored how AI—more specifically machine learning and reinforcement learning—can improve decision-making in supply chains, such as how to cut waste," added Paola Perez Tostado, a visiting student from University of the Americas Puebla.
Working with Nematollahi, Baik and Tostado received hands-on research experience and learned essential skills in operational management research. Over four months, they studied existing research, collected real-world data, built a computer model, and ran simulations. Their model included the variables that retailers must consider when selling perishable items, such as pricing, quality, and promotion strategies. Using this novel model, the team of three ran simulations to analyze how these variables could affect consumer demand for perishable items and, in turn, influence retailers’ strategies.
Discovering Mitacs: A door to global opportunities
Baik discovered the Mitacs GRI program through an online announcement.
"I saw the announcement on my university website," said Baik. "They were recruiting students for research internships in Canada, and I thought it would be a rare and meaningful opportunity. So, I applied. I had to pass an internal screening process at my school. After I got accepted, I had to fill out all the forms on the Mitacs intern homepage. It took a lot of paperwork, but it was totally worth it."
Tostado’s story is more serendipitous.
"I heard about the GRI program through my ex-roommate," said Tostado. "She had the opportunity to complete an internship in Montreal. When she came back, she told me all about it. I had no clue the program even existed before that!"
While they spent most of their time working away on their project at Edwards, they also got to explore Saskatoon’s gorgeous summer. They even got to visit Wanuskewin on National Indigenous Peoples Day!
Reflections and key takeaways from the Globalink Research Internship
Reflecting on their experience, the team agreed that the GRI program has made a positive impact on their careers.
"The program helped me see how I can combine the skills that I've learned in university to create real-world impacts," said Tostado. "It gave me a lot of insight into conducting research and more confidence to pursue a career path in this area or tackle interdisciplinary challenges. Data science and sustainability are not something that you see very often together and it's something I'm really interested in. I would really like to continue with graduate studies in this field in the future. And beyond that, it was meaningful. It was my first time learning about Indigenous cultures in Canada. My mom’s side of the family is connected to our Indigenous roots, so it gave me a chance to see parallels between both cultures and reflect on how important traditions are."
"I may not continue this exact project, but I would like to use the ideas and skills that I learned from this project to study shipping logistics," said Baik. "Shipping companies frequently change their prices, so I want to study how much they discount and when they decide to discount by using the reinforcement learning model that I learned from this experience."
For Nematollahi, the GRI program is a valuable asset to advance his research program as an early-career researcher.
"I first heard about Mitacs from the Office of the Vice President's Research at USask," said Nematollhi. "They invited Dr. Zsuzsa Papp, Senior Advisor, Business Development at Mitacs, to talk about Mitacs opportunities. It was the first time I heard about the GRI program, and it was really interesting. Getting to know these different opportunities is important, especially for early career researchers who are worried about funding or are deterred by the long and complex application processes in other opportunities. The GRI program has an easy application process and is rewarding."
Visit the Mitacs Globalink Research Award page to learn more about the program.