Name of File 5600-70-20.pdf cached at 17/12/2017 19:25:03 - with 2 pages. pdfPath: E:\k9.istructe.org\CMS\webtest\files\d5\d5edce74-01f8-467d-99ce-c804d7862d12.pdf. thumbPath: E:\k9.istructe.org\CMS\webtest\files\pdfthumbs\d5edce74-01f8-467d-99ce-c804d7862d12_1.png. objDoc: 1 - True. objPreview.Log: . strFileName: d5edce74-01f8-467d-99ce-c804d7862d12_1.png

Members/subscribers must be logged in to view this article

Verulam

Last words on the ‘genetic algorithm’ Last month (Verulam, 15 September) we attempted to summarise a long letter from Professor Jenkins in which he set out to expound, particularly in response to Stefan Tietz (21 July), the aims and potential achievements of his 'genetic algorith’ for structural optimisation. While we endeavoured to cover his main points, Professor Jenkins asks our indulgence in that he feeh we did not adequately present his answers to the ‘frank and uninhibited statements made by contributors’. His now proffered ‘closure’ is as follows: Mr Tietz has missed the point when he attributes functions to the GA that it does not possess. He questioned what he called ‘basic assumptions’ which did not appear in my paper. Nowhere do I talk about ‘rules’ for ‘perfection of design’ or for ‘construction costs’. I have been at pains to make the point that the GA is simply a tool in the designer’s toolkit. I am firm in my belief that structural design is an art practised by designers supported by scientific and engineering principles and methods, including computer-based methods. In this context I must urge Mr Tietz to take the opportunity to enlarge his definition of professional judgment. ‘The competent assessment of a huge number of variables’ is exactly what the GA does! Verulam