|
Post by Sunil Kumar on Aug 7, 2007 22:02:07 GMT -5
Under what conditions SORA would converge in single iteration?
One answer could be if the MPP remains fixed as design variable changes, then SORA would find an optimum which will not change in subsequent iterations. Hence, it has converged in one iteration.
But then question becomes under what conditions MPP will not move from one iteration to next?
|
|
liudu
New Member
Posts: 3
|
Post by liudu on Aug 8, 2007 12:13:15 GMT -5
When the performance function is linear, the MPP in standard normal space is fixed for any design point. If the performance function is not linear, one cal always find two designs, such that two MPPs in standard normal space are different.
|
|
|
Post by haftka on Aug 9, 2007 16:52:29 GMT -5
Linear in what? Random variables only? both random and design variables?
|
|
|
Post by sk99310 on Aug 15, 2007 23:33:35 GMT -5
If all random variables are linearly related in a constraint, and other design variables are non linear then SORA does not converge in single iteration. In fact, MPP changes in every iteration too.
what could be linear constraint problem in which SORA will converge in one iteration?
|
|
|
Post by Sunil Kumar on Aug 18, 2007 12:43:04 GMT -5
One answer to above problem is when there is only one random variable. Then MPP won't change from one iteration to another and then SORA will converge in one iteration.
|
|
|
Post by liuxin on Sept 13, 2015 7:07:46 GMT -5
It is really difficult for me to learn the SORA method. Is there any kind person like to share a matlab code for SORA ? I will really apreciate for that! Please send it to my email:liux23@126.com. Thank you very very much !!
With an eager mind to learn SORA.
|
|