[ Français ]

Thematic Semester: The Mathematics of Decision Making, January-June 2020

The overarching theme of this proposal is the understanding and optimization of large-scale systems, with an emphasis on computational issues. One of its aim is to illustrate how recent advances in articial intelligence (machine learning, for one) will expand the range of practical industrial and societal problems that optimization methods will have to address in the near future. Within this framework, the interplay between optimization and data science will be outlined. Actually, machine learning techniques can help to solve large-scale, ill-conditioned mathematical programs, and also feed them with features learned from a population of agents. Conversely, optimization algorithms, either exact or heuristic, are routinely used to train neural networks, for instance.

The program will cover a fairly wide array of topics that involve issues such as user behaviour, mixed continuous-discrete variables, randomness, all arising in large-scale, real-life complex systems. Their proper analysis requires knowledge in game theory and decomposition, which justies the integration of activities devoted to game theory and column generation, respectively. As a complement, two important applied areas, namely voting systems and healthcare, will complete the program.

The thematic program will be scheduled from January to June 2020, most events having a duration of one week.