2 February 2018, 12:00 - 14:00
ARITI is delighted to welcome Dr Daniel Karapetyan from University of Essex.
Prof Karapetyan's presentation followed by a Q&A session will take place 12-1pm. Lunch and networking will take place 1-2pm.
Much of the automated decision support relies on our ability to solve discrete optimisation problems. Due to the diversity of applications and, as a result, the diversity of optimisation problems, design of optimisation algorithms is a labour-intensive and time-consuming work requiring expensive human expertise. We will discuss a recent framework, called Conditional Markov Chain Search (CMCS), proposed to enable automated design of heuristic optimisation algorithms. We will show that CMCS is flexible enough to model several standard metaheuristics yet its behaviour is completely parameter-controlled. We will discuss questions related to generation of efficient CMCS configurations and show that our approach has already outperformed the state-of-the-art human-designed methods on two optimisation problems.
We will assume that the audience is familiar with the concept of discrete optimisation.
Please register via the Eventbrite link below. There is no need to print your ticket. Just turn up on the day.