BioMaths Colloquium Series - 2016/17
30 June 2017 - 3pm Maths Seminar Room
(room 224 Talbot Building 2nd floor)
Taming Nature Inspired Evolutionary Optimisation Algorithms
Dr Sean Walton
(Department of Computer Science, Swansea University, UK)
 |
| Image by Sean Walton |
Our last BioMaths Colloquium Series for the 2016/17 academic year will feature a seminar by Dr Sean Walton from the Department of Computer Science at Swansea University (UK). Jon is a computer science lecturer, interested in mesh generation and optimisation algorithms, evolutionary optimisation algorithms and intelligent design, web development, and computer games design - for the latter he even founded a company, Pill Bug Interactive. He is particularly interested in optimisation of computer design and in harnessing the power of evolution to do so. Sean is part of the visual computation research group and is passionate also about teaching, especially of web development.
Abstract
Abstract
Can a computer automatically design a car, a boat or an aircraft? This is a driving question at the heart of design optimisation. Increasingly researchers are looking towards nature for inspiration to achieve this goal, making a computer evolve a car instead of designing it.
By its nature evolution is hard to control, which causes some problems when trying to use it to do a specific task. What seems like a simple solution to a complex problem can quickly become difficult to use. The inherit randomness means you can never be completely certain you’ll get a good design quickly enough. This has led to complex adaptive algorithms which try to control mutation and survival rates from generation to generation to get the best design in the shortest time.
In this talk I will give a brief overview of evolutionary optimisation algorithms, discuss the problems with them and suggest some potential solutions.
By its nature evolution is hard to control, which causes some problems when trying to use it to do a specific task. What seems like a simple solution to a complex problem can quickly become difficult to use. The inherit randomness means you can never be completely certain you’ll get a good design quickly enough. This has led to complex adaptive algorithms which try to control mutation and survival rates from generation to generation to get the best design in the shortest time. In this talk I will give a brief overview of evolutionary optimisation algorithms, discuss the problems with them and suggest some potential solutions.
The discussions will continue over biscuits and tea/coffee after the seminar.
Hope to see many of you!
No comments:
Post a Comment