Mittagsseminar (in cooperation with A. Steger, D. Steurer and B. Sudakov)

Date and Time: Thursday, October 04, 2018, 12:15 pm

Duration: 30 minutes

Location: OAT S15/S16/S17

Speaker: Johannes Lengler

The univariate marginal distribution algorithm

Estimation-of-distribution algorithms (EDAs) are optimization algorithms which maintain a probability distribution over the search space: their belief of where good solutions are. In each round, they sample from their distribution and update it based on the fitness of the samples. Although EDAs have been a trend in black-box optimization in recent years, our theoretical understanding of them is quite limited, even for the simplest ones.
In this talk, I will present some recent joint results with Dirk Sugholt and Carsten Witt on the univariate marginal distribution algorithm. This algorithm operates on the hypercube, and its distribution is a product measure, i.e., it treats all bits independently. This simple algorithm shows a surprisingly complex performance landscape even on the simple OneMax function.

Upcoming talks     |     All previous talks     |     Talks by speaker     |     Upcoming talks in iCal format (beta version!)

Previous talks by year:   2023  2022  2021  2020  2019  2018  2017  2016  2015  2014  2013  2012  2011  2010  2009  2008  2007  2006  2005  2004  2003  2002  2001  2000  1999  1998  1997  1996  

Information for students and suggested topics for student talks

Automatic MiSe System Software Version 1.4803M   |   admin login