MATLAB Central Newsreader - mistake in poll method for pattern search?

mistake in poll method for pattern search?

Dave Brackett — Tue, 23 Sep 2008 00:50:17 -0400

I think I may have found a mistake in the 'PollMethod' option for Pattern Search. From the help for 'psoptimset', the possible options for the polling method are:

[ 'MADSPositiveBasisNp1' | 'MADSPositiveBasis2N' | 'GPSPositiveBasisNp1' | {'GPSPositiveBasis2N'} ]

However, when trying to set the option as 'GPSPositiveBasis2N' which is apparently the default anyway, I get this message;

Invalid value for OPTIONS parameter PollMethod: must be 'GPSPositiveBasisNp1', 'GPSPositiveBasisNp1',
'MADSPositiveBasisNp1', or , 'MADSPositiveBasis2N'.

Strangely 'GPSPositiveBasis2N' isn't even in this list so I can't set it as the polling method, while 'GPSPositiveBasisNp1' is listed twice. How come? My Matlab version is 2008a.

Re: mistake in poll method for pattern search?

Steven Lord — Tue, 23 Sep 2008 13:55:39 -0400

"Dave Brackett" <davebrackett@hotmail.com> wrote in message
news:gb9eg9$5bm$1@fred.mathworks.com...
>I think I may have found a mistake in the 'PollMethod' option for Pattern
>Search. From the help for 'psoptimset', the possible options for the
>polling method are:
>
> [ 'MADSPositiveBasisNp1' | 'MADSPositiveBasis2N' | 'GPSPositiveBasisNp1' |
> {'GPSPositiveBasis2N'} ]
>
>
> However, when trying to set the option as 'GPSPositiveBasis2N' which is
> apparently the default anyway, I get this message;
>
> Invalid value for OPTIONS parameter PollMethod: must be
> 'GPSPositiveBasisNp1', 'GPSPositiveBasisNp1',
> 'MADSPositiveBasisNp1', or , 'MADSPositiveBasis2N'.
>
>
> Strangely 'GPSPositiveBasis2N' isn't even in this list so I can't set it
> as the polling method, while 'GPSPositiveBasisNp1' is listed twice. How
> come? My Matlab version is 2008a.

I've tried to reproduce this behavior in Release 2008a but was not able to
do so:

psoptimset('PollMethod', 'GPSPositiveBasis2N')

returned the options structure I expected. Can you check to see if this
command works for you, and if it does what's different between it and the
call that errored?

I was able to confirm that there's a typo in the error message; it's been
reported to the development staff. One of the instances of
GPSPositiveBasisNp1 should be GPSPositiveBasis2N.

--
Steve Lord
slord@mathworks.com

Re: mistake in poll method for pattern search?

Dave Brackett — Tue, 23 Sep 2008 16:43:02 -0400

"Steven Lord" <slord@mathworks.com> wrote in message <gbasgr$efu$1@fred.mathworks.com>...
>
> "Dave Brackett" <davebrackett@hotmail.com> wrote in message
> news:gb9eg9$5bm$1@fred.mathworks.com...
> >I think I may have found a mistake in the 'PollMethod' option for Pattern
> >Search. From the help for 'psoptimset', the possible options for the
> >polling method are:
> >
> > [ 'MADSPositiveBasisNp1' | 'MADSPositiveBasis2N' | 'GPSPositiveBasisNp1' |
> > {'GPSPositiveBasis2N'} ]
> >
> >
> > However, when trying to set the option as 'GPSPositiveBasis2N' which is
> > apparently the default anyway, I get this message;
> >
> > Invalid value for OPTIONS parameter PollMethod: must be
> > 'GPSPositiveBasisNp1', 'GPSPositiveBasisNp1',
> > 'MADSPositiveBasisNp1', or , 'MADSPositiveBasis2N'.
> >
> >
> > Strangely 'GPSPositiveBasis2N' isn't even in this list so I can't set it
> > as the polling method, while 'GPSPositiveBasisNp1' is listed twice. How
> > come? My Matlab version is 2008a.
>
> I've tried to reproduce this behavior in Release 2008a but was not able to
> do so:
>
>
> psoptimset('PollMethod', 'GPSPositiveBasis2N')
>
>
> returned the options structure I expected. Can you check to see if this
> command works for you, and if it does what's different between it and the
> call that errored?
>
> I was able to confirm that there's a typo in the error message; it's been
> reported to the development staff. One of the instances of
> GPSPositiveBasisNp1 should be GPSPositiveBasis2N.
>
> --
> Steve Lord
> slord@mathworks.com
>

I actually found that the reason it was throwing up the error was because I had the poll method option stored within a cell. Once I had extracted the characters from the cell using char() it worked. It was the typo in the error message that was confusing me. Thanks.