Description
Title: Mathematical foundations of randomised optimisation algorithms.Abstract: Randomised Optimisation Algorithms such as evolutionary algorithms, simulated annealing or estimation of distribution algorithms implement a general idea of how to search for solutions for (hard) optimisation problems. They iteratively sample candidate solutions from a search space and assess the quality of a solution using an objective function. They provide a powerful and flexible way of tackling different complex problems where classical optimisation methods fail. While the general idea is to apply such algorithms 'right out of the box', in practice it is almost always necessary to adjust them to the problem at hand by modifying the overall search strategy to achieve acceptable performance. It is thus highly desirable to obtain a clear understanding of the working principles of different operators and strategies. Mathematical analysis can provide such an understanding, including properties of problems and operators, parameterisation, and limitations of different approaches, and can inspire the design of better algorithms. Over the last few decades significant progress on mathematical foundations of Randomised Optimisation Algorithms has been made. This talk will provide an overview of the main lines of research in the area with a focus on runtime and anytime analysis in combinatorial optimisation. It will highlight example results and illustrate how these results can be used for the modification and development of algorithms in relevant applications. I will also point out future research directions with the aim to initiate a dialogue between researchers interested in theory and applications.
| Period | 12 Dec 2023 |
|---|---|
| Event title | 23rd International Conference on Intelligent Systems Design and Applications |
| Event type | Conference |
| Degree of Recognition | International |