Comments on: How to make a straight line square http://www.husdal.com/2000/09/01/how-to-make-a-straight-line-square/ Journal articles and papers, books and book chapters, research reports and whitepapers, blogs and websites Sat, 28 Jan 2012 19:44:00 +0000 hourly 1 http://wordpress.org/?v= By: Jan Husdalhttp://www.husdal.com/2000/09/01/how-to-make-a-straight-line-square/#comment-1557 Jan Husdal Mon, 13 Jul 2009 06:13:16 +0000 http://transportationgis.wordpress.com/2000/09/01/how-to-make-a-straight-line-square/#comment-1557 Thanks for your compliments on my now 10-year-old thesis. I am glad that it is still of some use. As to the figure, that is not my MY figure but taken from the reference mentioned. I did make the same calculations as that author and I did not see the 'flaw'. Thanks for pointing that out. Thanks for your compliments on my now 10-year-old thesis. I am glad that it is still of some use. As to the figure, that is not my MY figure but taken from the reference mentioned. I did make the same calculations as that author and I did not see the ‘flaw’. Thanks for pointing that out.

]]> By: Irmela Herzoghttp://www.husdal.com/2000/09/01/how-to-make-a-straight-line-square/#comment-1555 Irmela Herzog Mon, 13 Jul 2009 05:54:12 +0000 http://transportationgis.wordpress.com/2000/09/01/how-to-make-a-straight-line-square/#comment-1555 The thesis is a very good overview over the different aspects of least-cost path calculations. But trying to understand the example of Fig. 3-9 I think I found a mistake: On the green path, the accumulated costs are 0 -> 2.1 -> 4.1 (!!!) -> 6.1, on the red one: 0 -> 1 -> 2.5 -> 5.3, and therefore the red path is optimal. The argument that steepest descent is not optimal is not invalidated by this mistake. The full implementation of Dijkstra's algorithm includes backlinks so that the optimal path is found - always! Best wishes Irmela The thesis is a very good overview over the different aspects of least-cost path calculations. But trying to understand the example of Fig. 3-9 I think I found a mistake: On the green path, the accumulated costs are 0 -> 2.1 -> 4.1 (!!!) -> 6.1, on the red one: 0 -> 1 -> 2.5 -> 5.3, and therefore the red path is optimal.
The argument that steepest descent is not optimal is not invalidated by this mistake. The full implementation of Dijkstra’s algorithm includes backlinks so that the optimal path is found – always!

Best wishes
Irmela

]]>