User talk:Jonmsmith

From Wikipedia, the free encyclopedia

Welcome!

Hello, Jonmsmith, and welcome to Wikipedia! Thank you for your contributions. I hope you like the place and decide to stay. Here are some pages that you might find helpful:

I hope you enjoy editing here and being a Wikipedian! Please sign your name on talk pages using four tildes (~~~~); this will automatically produce your name and the date. If you need help, check out Wikipedia:Questions, ask me on my talk page, or ask your question and then place {{helpme}} after the question on your talk page. Again, welcome!  -- RHaworth 06:39, 15 April 2007 (UTC)

[edit] Bolonkins

I have replied about program management on my talk page. Sorry if I have been a bit harsh. After that, you may find the following request a bit cheeky, but I will ask it!

We recently had a serious attack of Bolonkins on Wikipedia - a glance at this AfD will probably tell you what I mean! I would be fascinated to know what the unofficial (or official) view from inside NASA is of Alexander Bolonkin and his designs. Are any of them feasible? Are any of them notable (ie. do they warrant Wikipedia articles)? -- RHaworth 06:39, 15 April 2007 (UTC)

[edit] T-integration

The claim that picking P = 3/2 in T-integration yields the Adams method has repeatedly been inserted, last by you. I find it very hard to believe that this is true. Could you please give a reference to a proof of this fact? Thanks, Jitse Niesen (talk) 08:47, 15 April 2007 (UTC)

Thanks for your kind reply. I'll get your book from the library and see whether that helps me understand what's going on. -- Jitse Niesen (talk) 09:26, 16 April 2007 (UTC)

Once you convince yourself, will you put the material on Adams method back in my contribution?jonmsmith 02:35, 17 April 2007 (UTC)

Sorry for the long delay. I tried to read your book Mathematical Modeling and Digital Simulation for Engineers and Scientists. It's tough going for me because it's written in the language of system and control theory with which I'm not very familiar, so I may well misunderstand some parts.
Anyway, equation (7-10) in the book is
X_n = X_{n-1} + \lambda T ( \gamma \dot{X}_n + (1-\gamma) \dot{X}_{n-1} ).
This seems to be the same as the formula in the article on T-integration. Using a gain of λ = 1 and a phase of γ = 3 / 2 yields
X_n = X_{n-1} + T ( \tfrac32 \dot{X}_n - \tfrac12 \dot{X}_{n-1} ),
which is the formula you wrote on my talk page. However, the Adams method — more specifically, the second-order explicit Adams(-Bashforth) method — is
X_n = X_{n-1} + T ( \tfrac32 \dot{X}_{n-1} - \tfrac12 \dot{X}_{n-2} );
this is equation (7-4) in the book. That looks quite different to me. So, I'm afraid I'm not convinced yet. -- Jitse Niesen (talk) 13:27, 19 May 2007 (UTC)

jonmsmith 23:10, 22 May 2007 (UTC)Jitse... Good to hear from you. I always enjoy exchanges on numerical analysis. Please take a moment and read pages 258 thru 260 of my book (in library) Mathematical Modeling and Digital Simulation for Engineers and Scientists 2nd edition. Note that the distinction between using a numerical integrator implicitly vs explicitly.

The form of the Adams integrator you derived is the explicite form. Used directly as a numerical algorithm in a digital computer. This version has a built in destabalizing delay because the present is computed in terms of the past.

The implicite form of the Adams Integrator is the one I propose to include in my Wiki article. It computes the present in terms of the present. That leads to an implicite difference equation that has to be solved for the present in terms of the past. It is done by substituting the numerical integrator into the differential equation and the implicite difference equation solved for present in terms of past algebraically. The resulting difference equation being the algebraically integrated differential equation.

Note that all numerical integrators can be used either implicitly or explicitly. The difference is the resuling difference equation that integrates the differential equation. The implicite versions are the more accurate. The version I use is the implicite version of the Adams integrator, again see my book.

If you have never used integrators this way, Try it with the simple euler integrator (present in terms of past) and the rectangular integrator (present in terms of the present) to see the difference. It is a great insight if you have never done it before.

What do you think? Can I include the implicite Adams in my article?

jonmsmith 01:27, 22 July 2007 (UTC) Jitse. Are you ok with my including implicite Adams in my article? July 21 2007.

jonmsmith 23:10, 22 May 2007 (UTC)Jon Michael Smith

Okay, I think I now see what you're doing. You take the explicit Adams method
X_n = X_{n-1} + T ( \tfrac32 \dot{X}_{n-1} - \tfrac12 \dot{X}_{n-2} );
and then say that you're converting it to implicit form:
X_n = X_{n-1} + T ( \tfrac32 \dot{X}_{n} - \tfrac12 \dot{X}_{n-1} ).
Well, I guess you could, but that's a different method. It's not a well-known technique, as the article used to say. In fact, I've never seen it used. I think it's very misleading to call it an Adams method. There is a method called the "implicit Adams method", also known as the "Adams-Moulton method". In second-order form, it is the trapezoidal rule. -- Jitse Niesen (talk) 12:46, 5 September 2007 (UTC)

[edit] April 2008

Hi, the recent edit you made to T-integration has been reverted, as it appears to be unconstructive. Use the sandbox for testing; if you believe the edit was constructive, ensure that you provide an informative edit summary. You may also wish to read the introduction to editing. Thanks. Calvin 1998 (t-c) 01:48, 15 April 2008 (UTC)