Presentation Information
[8p-S202-4]Revolutionizing through the Synergy of Machine Learning and Mathematical Optimization (MOAI)
〇Mikio Kubo1,2 (1.Tokyo Univ. of Marine Sci. and Tech., 2.MOAI Lab.)
Keywords:
MOAI
In recent years, research integrating Artificial Intelligence (AI) with Mathematical Optimization (MO), a field we term MOAI, has rapidly advanced. This presentation focuses on two key MOAI approaches considered effective for real-world optimization problems.
The first is a new paradigm for rapidly solving NP-hard real-world problems. Real-world scenarios often involve repeatedly solving similar problem instances daily. In such an online environment where similar instances continuously arise, various strategies can overcome NP-hardness.
The second approach involves using Agentic AI as an alternative to traditional optimization consulting. MO projects for real-world problems typically involve iterative client hearings, model extraction, optimization algorithm selection, implementation, and testing, processes that are usually very time-consuming and costly. We propose a methodology to expedite this process using Agentic AI.
The first is a new paradigm for rapidly solving NP-hard real-world problems. Real-world scenarios often involve repeatedly solving similar problem instances daily. In such an online environment where similar instances continuously arise, various strategies can overcome NP-hardness.
The second approach involves using Agentic AI as an alternative to traditional optimization consulting. MO projects for real-world problems typically involve iterative client hearings, model extraction, optimization algorithm selection, implementation, and testing, processes that are usually very time-consuming and costly. We propose a methodology to expedite this process using Agentic AI.