• generate 100 random number or use the test data provided by the standard; • write Convolution Encoder using MATLAB script; Then you need to verify the Convolution Encoder by • calling MATLAB built in Convolution Encoder function ; • calling MATLAB built in Viterbi Decoder function

0 |

12/8/2009 9:25:20 PM

The command you need is "doc" and if that doesn't make sense try "why"

Jan Simon wrote: >> I still have all of my Dungeons & Dragons dice. Except that I never >> did figure out how to synthesize a fair D17 (17 sized die.) > Use a gyroskop (spinning top) with the ring cut to a regular polygon > with 17 sides. The side the gyroskop stops on is the wanted random number. Interesting... it turns out that there -is- a regular (compass and straight-edge) construction for the 17-gon and the 257-gon, so what you propose would be possible with a fair die -- though it might be a bit tricky to find a material that will hold 17 vertices without chipping. http://en.wikipedia.org/wiki/Heptadecagon The gyroscope part could be eliminated by taking two 17-sided (plus a base) pyramids and gluing them together at the base, and then using the same number on exactly two faces.

0 |

1/20/2010 8:22:48 PM

Walter Roberson <roberson@hushmail.com> wrote in message <4B563E41.7080509@hushmail.com>... > Jan Simon wrote: > > > Does "generate 100 random number" allow forwarding this task to RAND as > > generator, or does this mean that the person has to generate them > > personally? I've read that humen choose 2, 7, 13 and 10 too seldom, > > because they are too odd or too even. > > We weren't told of any requirement for uniform random distribution ;-) > > I still have all of my Dungeons & Dragons dice. Except that I never did figure > out how to synthesize a fair D17 (17 sized die.) You can try to modify this one: http://gamesbyemail.com/DiceGenerator Jos

0 |

1/21/2010 7:34:04 AM

"Jan Simon" <matlab.THIS_YEAR@nMINUSsimon.de> wrote in message <hj5ih3$kv8$1@fred.mathworks.com>... > Dear Walter! > > > I still have all of my Dungeons & Dragons dice. Except that I never did figure > > out how to synthesize a fair D17 (17 sized die.) > > Use a gyroskop (spinning top) with the ring cut to a regular polygon with 17 sides. The side the gyroskop stops on is the wanted random number. Just being curious: Is there a general algorithm to get random, uniformly distributed integers between 1 and N when all you have is an M-sided dice? Jos

0 |

1/21/2010 8:07:03 AM

On Jan 19, 5:46=A0pm, "Rob Campbell" <mat...@robertREMOVEcampbell.removethis.co.uk> wrote: > The command you need is "doc" and if that doesn't make sense try "why" ---------------------------------------------------------- As a fun aside, go to Google, and if you have the suggestions feature turned on, type in "why." It's very funny what prior questions show up as suggestions!

0 |

1/21/2010 11:16:13 AM

"Jos (10584) " <#10584@fileexchange.com> wrote in message <hj91v7$5rg$1@fred.mathworks.com>... > "Jan Simon" <matlab.THIS_YEAR@nMINUSsimon.de> wrote in message <hj5ih3$kv8$1@fred.mathworks.com>... > > Dear Walter! > > > > > I still have all of my Dungeons & Dragons dice. Except that I never did figure > > > out how to synthesize a fair D17 (17 sized die.) > > > > Use a gyroskop (spinning top) with the ring cut to a regular polygon with 17 sides. The side the gyroskop stops on is the wanted random number. > > Just being curious: > > Is there a general algorithm to get random, uniformly distributed integers between 1 and N when all you have is an M-sided dice? > > Jos Sure. Why not? Just change bases. An M-sided die allows you to work in base M, or if M is a composite number, to work with more general radix numbers. Now I wish to generate a random integer in base N. Just generate enough "digits" in base M, rejecting those that fall outside the desired range when converted to base N digits. However, I assume that your real question is one where no rejection is necessary, therefore no super sampling. I'll argue that a rejection scheme can always be made quite efficient, with a limited amount of resampling for large enough sample sizes. John

0 |

1/21/2010 11:55:21 AM

"John D'Errico" <woodchips@rochester.rr.com> wrote in message <hj9fb9$b16$1@fred.mathworks.com>... > "Jos (10584) " <#10584@fileexchange.com> wrote in message <hj91v7$5rg$1@fred.mathworks.com>... > > "Jan Simon" <matlab.THIS_YEAR@nMINUSsimon.de> wrote in message <hj5ih3$kv8$1@fred.mathworks.com>... > > > Dear Walter! > > > > > > > I still have all of my Dungeons & Dragons dice. Except that I never did figure > > > > out how to synthesize a fair D17 (17 sized die.) > > > > > > Use a gyroskop (spinning top) with the ring cut to a regular polygon with 17 sides. The side the gyroskop stops on is the wanted random number. > > > > Just being curious: > > > > Is there a general algorithm to get random, uniformly distributed integers between 1 and N when all you have is an M-sided dice? > > > > Jos > > Sure. Why not? Just change bases. > > An M-sided die allows you to work in base M, > or if M is a composite number, to work with > more general radix numbers. > > Now I wish to generate a random integer in > base N. Just generate enough "digits" in base > M, rejecting those that fall outside the desired > range when converted to base N digits. > > However, I assume that your real question is > one where no rejection is necessary, therefore > no super sampling. I'll argue that a rejection > scheme can always be made quite efficient, > with a limited amount of resampling for large > enough sample sizes. > > John Thanks John for this answer. Your assumption is correct, however. I was thinking about some remapping of a few random numbers between 1 and N to the range 1 to M, without rejection or re-rolling the dice. And, yes, for some cases this is trivial (e.g., N=6, M =3). Jos

0 |

1/21/2010 2:47:04 PM

Jos (10584) wrote: > Just being curious: > Is there a general algorithm to get random, uniformly distributed > integers between 1 and N when all you have is an M-sided dice? If M and N are relatively prime, then you cannot get exact fairness, but you can get fairness to any desired (finite) degree of accuracy. Start with 0 to N-1, and using rational fractions, subdivide the interval into M parts. Roll the die to select which subrange of the interval you get. If that subrange contains only a single one of the original 1-N slots, then the selected number is that label. If the subrange spans multiple of the original 1-N slots, arithmetically subdivide the interval into M subintervals, roll the die, select the appropriate sub-interval... This process could, in theory, take forever, such as would be necessary to resolve -exactly- 1/3 into binary (a 2 sided die). The -expected- number of rolls would probably not be too bad to calculate, though the formula doesn't immediately spring to mind. But because it could take forever, if you want a guaranteed time-limit of T steps, you could be inaccurate to N / (M^T) (or perhaps T+1 instead of T, I'd have to work it through.) As fair as you have time for -- and no rejections.

0 |

1/21/2010 11:52:04 PM