Reply by Vladimir Vassilevsky February 21, 20082008-02-21

PARTICLEREDDY (STRAYDOG) wrote:
> well all dsp members, > i want to clarify under what > circumstances does the optimal and sub-optimal classifcations of a > particular algorithm is done.
If the algorithm approaches the theoretical limit, then it is called optimal. Otherwise it is suboptimal. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Reply by Rune Allnor February 21, 20082008-02-21
On 21 Feb, 13:33, "PARTICLEREDDY (STRAYDOG)" <particlere...@gmail.com>
wrote:
> well all dsp members, > &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295; &#4294967295;i want to clarify under what > circumstances does the optimal and sub-optimal classifcations of a > particular algorithm is done.
...
> ok, above is fine..now..on what basis (or for that matter, criterion) > algorithms are classfied into optimal and suboptimal
The main issue is whether there exists a limit to what performance is possible. If such a limit can be established[*] an algorithm is 'optimal' if its performance reaches the specified limit and 'suboptimal' if it doesn't. [*] Of course, one needs to establish exactly what limit one is talking about before it makes sense to use a term like 'optimal.' One could be talking about the Cramer Rao Bound, run-time, memory footprint, economical investments required... Different solutions optimized with respect to different performance goals would end up looking very different. Rune
Reply by PARTICLEREDDY (STRAYDOG) February 21, 20082008-02-21
well all dsp members,
                               i want to clarify under what
circumstances does the optimal and sub-optimal classifcations of a
particular algorithm is done. we say EQUALIZATION with error control
coding can be done in two ways

1. optimal receiver has to tackle out both ISI and error correcting
code be decoded jointly(i mean joint decoding mechanism)
2. suboptimal receiver has to tacke out equalization (related to above
ISI ) and the decoding (related to error correcting code) seperately..

ok, above is fine..now..on what basis (or for that matter, criterion)
algorithms are classfied into optimal and suboptimal

(A).  Is it based on number of functionalities tackled (say in above
case optimal has to tackle out both)
(B).  Or is it based on number of complexity of implementation (in
terms of number of computations)
(C).  Or is it COMBINATION of both (A) and (B)

thanks for all answers

thanks again in advance

particle (filter) reddy