Although established relatively late, the idea of operations research has a long history. The ancient Chinese legend of "Tian Ji horse racing" and the Greek Peloponnesian War in which Athens and Sparta each leveraged their own advantages to achieve victory are concrete manifestations of the core idea of operations research, "optimization". Operations research was born out of war and has entered modern times to serve various aspects of human life, including clothing, food, housing, and transportation. What are the concepts and applications of operations research? What is the relationship with another important discipline, artificial intelligence? International renowned operations researcher, the first Chinese recipient of the von Neumann Theory Award, and tenured professor at Stanford University, Ye Yinyu, gave an interview to answer the above questions. The summary of the interview transcript is as follows: Reporter: As a science, Operations Research (OR) has a short history and is somewhat distant from the daily lives of the general public. What is operations research? Why was operations research born? What is the purpose of establishing operations research? Ye Yinyu: From a historical perspective, the concept of "operations research" originated from World War II. In this largest war in human history, the implementation of each campaign requires the deployment of massive manpower and resources, as well as careful military decisions. At the macro level, there is a strategic level, while at the micro level, there is an operational level. This is the focus of operations research and the reason for its birth. With the continuous development of productivity and the advancement of science and technology, the previous "qualitative" methods are no longer sufficient to solve current problems. Only by "quantifying" problems can solutions be found. When individuals are unable to make decisions quickly and accurately in such a rapidly changing environment, the quantitative methods provided by operations research can help humans make better decisions. For example, in the developed financial field, quantitative finance and quantitative trading have emerged, indicating that the whole world is moving towards a more micro, detailed, and quantitative direction. The West often says "Devil is in the details", which is translated into Chinese as "details determine success or failure". When China introduced this science in the last century, it translated "Operations Research" as "Operations Research". I think this translation is good, but it has not fully reflected the meaning of "quantification" behind it. Ultimately, "optimization" is the core idea and purpose of operations research, which means obtaining the optimal solution in the process of solving a problem or achieving a goal. As early as the 1960s, mathematicians in China advocated using the idea of "optimization" to solve practical problems. One of them was the famous mathematician Mr. Hua Luogeng, whose "0.618 method" could find the most reasonable process conditions through fewer experiments, providing great help in improving industrial production efficiency at that time. On October 31, 2019, visitors to the "Traveling Far and Dreaming Together - Overseas Chinese and New China Special Exhibition" at the National Museum of China visited the manuscript of Hua Luogeng's "Starting from the Circle of the Unit" from the 1950s. Reporter Jia Tianyong from China News Service: What are the applications of operations research in daily life? Ye Yinyu: Operations research has a wide range of applications in daily life. Like the navigation software commonly used by everyone, satellite positioning can only provide real-time location information, but operations research is the one that plans reasonable travel routes for users. In modern industrial production, productivity has greatly developed and output has greatly increased. At the same time, there is an increasing demand for personalization and customization. Reasonably arranging industrial production is also a manifestation of operations research thinking. In the field of logistics, due to uncontrollable factors such as weather changes and road maintenance in reality, logistics transportation often cannot be carried out according to ideal planning. This requires the use of operations research to establish reasonable storage points and transportation lines to ensure the tight connection of various links in the supply chain. And the flights and railways that people travel on in their daily lives involve complex route and railway planning, requiring the planning of hundreds or thousands of flight and railway schedules in a day, all of which require operations research to solve one by one. On April 27, 2023, in Fuzhou, Fujian, at the 6th Digital China Construction Achievement Exhibition, the Beidou networking satellite model was showcased in the National Beidou Major Project Achievement Exhibition Area, attracting visitors. Photo by Lv Ming, China News Service reporter: "Solvers" are very important for solving operations research problems, and their focus is on optimization problems. The successful development of China's first large-scale operations research optimization algorithm solver is thanks to your years of guidance. What is the current development status of this field in China? What is the outlook? Ye Yinyu: The algorithm designed by operations research for optimization problems, implemented on a computer, is called a "solver". The West has long dominated the solver market, such as Xpress from FICO in Europe (later acquired by the United States) and Cplex from IBM in the United States, both of which are representative works of solvers. Chinese companies that needed to use solvers needed to purchase Western solver products because there were no developers using solvers in the country at that time. In recent years, due to changes in the external environment, some domestic enterprises have found it difficult to continue using Western solver products and must turn to "self-reliance". Some companies have approached me and my students, hoping that we can overcome unprecedented difficulties and develop China's own solvers. We have indeed achieved it. Does China's development of its own solver mean breaking the Western monopoly? I can say it is, but I don't think that's the point. Our goal is not just to break the "monopoly" or to create domestic substitutes for some Western products. We must strive to do our best and ensure that our products are used worldwide in the future. At present, we have indeed achieved many results in the development of solvers, some of which are world leading. Reporter: What is the relationship between operations research and artificial intelligence (AI)? Ye Yinyu: Some of my personal experiences can answer these questions. When I arrived in the United States in early 1982, my two academic pursuits were artificial intelligence (AI) and operations research (OR). At that time, my mentor assigned us a task to build an expert system for traditional Chinese medicine, which included information from experts and diagnostic methods for traditional Chinese medicine. At that time, the Internet had not yet appeared, and building such a system could only rely on continuous searching. We have encountered many problems, the most prominent of which is that some traditional Chinese medicine diagnostic methods have a very vague definition of "quantity", such as the appearance of many words such as "appropriate amount" and "small amount". I think this may be due to experience. In the minds of experts, words such as "moderation" should be the result of "quantification", but for outsiders, they are elusive. So under the conditions at that time, constructing such a system, data was not enough. But it was precisely these problems that sparked my interest in "quantification" and led me to pursue research in operations research. The currently popular artificial intelligence technologies, such as some large models, use the idea of operations research to train them as an optimization process. The optimization process of operations research can be achieved through advanced technologies such as artificial intelligence. Reporter: What are the advantages of operations research and artificial intelligence respectively? Ye Yinyu: I think both have their own strengths. Compared to artificial intelligence, the significant advantage of operations research is that designing an algorithm that does not require real data to participate can abstract problems and data without affecting the optimization process of operations research. Just like putting items of different values and types in a backpack, artificial intelligence needs to understand the specific nature and value of these items in order to provide solutions. Operations research optimization, on the other hand, provides an optimal algorithm that can find the best solution based on the algorithm, no matter what kind of items are put in. Simply put, it means adapting to changes with invariance and using "one" to "infinity". Compared to operations research, the development of artificial intelligence in recent years can be described as "flying into the homes of ordinary people", known to everyone. When using large models, people can ask questions in plain language and quickly come up with reasonable answers. However, operations research is a science that is not easily understood by the general public, but we are not familiar with it in our daily lives. On February 2, 2025, the robot band performed music at the Beijing Haidian New Year Science and Technology Temple Fair. Photo by Chen Fangting, China News Service reporter: Recently, the Nobel Prize in Physics and Chemistry were awarded to experts in the study of artificial intelligence. Some opinions on the internet believe that disciplines such as physics and chemistry are "about to disappear" under the impact of artificial intelligence. What is the relationship between artificial intelligence and fundamental disciplines? How should we view artificial intelligence? Ye Yinyu: These statements are exaggerated. I think it is precisely the result of the Nobel Prize that proves that the achievements of artificial intelligence are based on disciplines such as physics and chemistry. For example, this year's Nobel Prize in Physics is for research on artificial neural networks and machine learning, where "network" is a concept related to mathematics and physics. The emergence of artificial intelligence ultimately aims to solve practical problems in people's daily lives, and will not become a "castle in the air" detached from reality. These practical problems are related to fundamental sciences such as physics and mathematics. The fact that some teams in the field of artificial intelligence research and development that I have come into contact with have struggled to achieve tangible results in recent years proves this point. In addition, the implementation and development of artificial intelligence technology also require consideration of the question of "where it comes from and where it is going", which actually depends on the development of various disciplines, from multiple dimensions such as its origin, mechanism, and future. As a tool, the development of artificial intelligence itself is not the goal, solving practical problems and serving humanity is the goal. Reporter: What are the potential problems in the development of artificial intelligence? Ye Yinyu: In my opinion, the future development of artificial intelligence mainly depends on who will use this technology and how it will be used. Many people in the West worship artificial intelligence as a highly intelligent 'god', and I think this idea is problematic. We are developing artificial intelligence technology not to pursue the creation of a highly intelligent 'god'. If developers hold this idea, it will inevitably lead to social 'unfairness' - there are differences between people, and' intelligence 'should not be the standard for judging a person's value. In addition, there are also issues of energy consumption and time consumption in the development of artificial intelligence, especially energy consumption, which can bring about a large number of environmental problems. To solve the energy and time consumption of artificial intelligence, operations research methods are needed to help the development of artificial intelligence achieve twice the result with half the effort, save time and resources by optimizing the training process. The development of artificial intelligence is a common aspiration between China and the West, and environmental issues are a common problem faced by China, the West, and even the world. Therefore, I believe that there is a lot of cooperation between China and the West in this regard. (Wang Bingzhe, a doctoral candidate from Hillhouse School of Artificial Intelligence of Renmin University of China, contributed to this article) Interviewee profile: Yinyu Ye, a professor of Li Guoding in the Department of Management Science and Engineering and the Institute of Computational Mathematics Engineering of Stanford University, is one of the pioneers in the fields of interior point optimization algorithms, online linear planning and learning, reinforcement learning, and Markov process algorithm analysis. He has won the 2009 John von Neumann Theory Award, the 2012 International Mathematical Programming Conference (ISMP) Tseng Lecturer Award, and the 2014 American Society for Applied Mathematics Optimization Award. (New Society)