generate double persision random value
in [C++]
|
From: sposes on 9 Dec 2005 16:23 After searching to no evail for an answer to generate double presision random values I come here for assistance. I am still very new to C++ and have a situation where I need to generate a radom number that contains a fractional part. I have the following statement: currentValue = minValue + (rand() % maxValue); where currentValue, minValue and MaxValue are all of type double. As it is there is a compile error that states invalid operands of types 'int' and 'double' to binary 'operator%' I think this means that % requires an int on both sides, but if this is ture how am i to solve this delema. Im guessing that casting somehow might help me, but I am not very familiar with this. If I were to cast maxValue as an int, this might satisfy the % operator problem but prevents currentValue from ever achiveing a real value such as 3.56 Can someone please direct me on how to aproach this problem. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ]
From: Carlos Moreno on 9 Dec 2005 21:02 sposes wrote: > After searching to no evail for an answer to generate double presision > random values I come here for assistance. I am still very new to C++ > and have a situation where I need to generate a radom number that > contains a fractional part. I have the following statement: > > currentValue = minValue + (rand() % maxValue); > > where currentValue, minValue and MaxValue are all of type double. As > it is there is a compile error that states invalid operands of types > 'int' and 'double' to binary 'operator%' > > I think this means that % requires an int on both sides, but if this is > ture how am i to solve this delema. Im guessing that casting somehow > might help me, but I am not very familiar with this. If I were to cast > maxValue as an int, this might satisfy the % operator problem but > prevents currentValue from ever achiveing a real value such as 3.56 > > Can someone please direct me on how to aproach this problem. 1) If you happen to be working on a Unix/Linux system, they typically provide a rand48 family of functions which, in addition to being usually higher quality than the standard rand() function, it does include a drand48() function that returns a double-precision floating point between 0 and 1 (including 0, excluding 1). With this, you can simply use some math to come up with a linear (or rather, affine) formula to map the range [0,1) to your given range. 2) If you're restricted to the standard rand() function, perhaps the second best option would be to map the entire [0,RAND_MAX] range to your range *of floating point*. This is actually bad, since you're still restricted to only 4.2 billion possible values, even if your range contains a higher amount of representable values. 3) Have you checked Boost? (http://www.boost.org) IIRC, they have several high-quality random numbers generators. As usual, when asking about random numbers, the necessary question in response to your question is: what do you need them for? Depending on the application, the criteria to select "the right thing to do" can vary quite wildly. HTH, Carlos -- [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ]
From: peter steiner on 9 Dec 2005 21:03 sposes wrote: > After searching to no evail for an answer to generate double presision > random values I come here for assistance. I am still very new to C++ > and have a situation where I need to generate a radom number that > contains a fractional part. I have the following statement: > > currentValue = minValue + (rand() % maxValue); > > where currentValue, minValue and MaxValue are all of type double. As > it is there is a compile error that states invalid operands of types > 'int' and 'double' to binary 'operator%' > > I think this means that % requires an int on both sides, but if this is > ture how am i to solve this delema. Im guessing that casting somehow > might help me, but I am not very familiar with this. If I were to cast > maxValue as an int, this might satisfy the % operator problem but > prevents currentValue from ever achiveing a real value such as 3.56 > > Can someone please direct me on how to aproach this problem. if you want random floating point values you don't need to use the modulo operator. rather calculate a value between 0 and 1 with the help of <cstdlib> RAND_MAX and the use that to get a value inside your specified range: double drand(double min, double max) { double r = static_cast<double>(rand()) / RAND_MAX; // [0..1] return min + (max - min) * r; } -- peter [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ]
From: gerg on 9 Dec 2005 21:18 here is a link if you need a fine level of granularity: http://web.comlab.ox.ac.uk/oucl/work/richard.brent/random.html in general just do something like this double rand_val = ((double)rand()) / ((double)MAX_RAND); double val = (double)minval + rand_va*((double)maxval-(double)minval); rand() has a real crappy range{0..32767} so using something else is a good idea. [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ]
From: Carlos Moreno on 10 Dec 2005 07:58
gerg wrote: > rand() has a real crappy range{0..32767} so using something else is a > good idea. Errr, no. (I mean, yes, rand() is crappy, but for other reasons :-)) RAND_MAX is implementation-dependent -- on my platform, it is 2^31 - 1 (approx. 2.1 billion) Carlos -- [ See http://www.gotw.ca/resources/clcm.htm for info about ] [ comp.lang.c++.moderated. First time posters: Do this! ] |