I've generated a series of random numbers from a known seed in C using srand() and rand() from stdlib. I now need to generate the same series of numbers using the same seed from C in Java.
Java's Random class documentation states it uses a "linear congruential formula". The documentation I've found on rand() says it uses a "linear congruential" generator, although I'm not sure if this is for one specific implementation.
Does anyone know if both generators will produce the same numbers if given the same seed or if a port of srand() and rand() exists for Java?
The C standard does not dictate the implementations of srand() and rand(). As such, different environments (OS, C libraries, architecture, etc.) will more than likely produce sequences of numbers that are different for the same seed value.
Also, the implementation Java's Random class is not bound to any particular algorithm. Here again, different JVMs may very well produce different sequences for the same seed value. In addition, the implementation will more than likely not be tied to the standard C functions. Which means, the Java produced sequence will be different than a C sequence using the same seed.
If you truly need to generate random sequence in Java to match exactly that of the standard C functions, the best you could hope to do is replicate the sequence for a particular environment. This would require creating a JNI library to proxy calls to srand() and rand() or creating some other external process that makes the calls and is invoked from Java. Either way, that's a lot of complexity and more program maintenance.
If in fact all you need are random sequences that appear to be uniformly distributed, regardless of the exact values, then use Random as is. It is more than sufficient for most RNG needs.
As said in other answer, the C standard doesn't even specify that rand() will return the same sequence across different C platforms (libraries), and likewise nothing in Java guarantees that it matches any given C (or other Java) implementation. You could use JNI to call the specific C implementation on that platform, but this would only guarantee the same sequence to be produced when both the C and Java programs are run on the same platform using the same C library.
If you wish to ensure the same sequence in all situations, you need to implement the same random number generator in both languages. A simple example can be found in POSIX.1-2001, and is quoted on many man 3 rand pages:
static unsigned long next = 1;
/* RAND_MAX assumed to be 32767 */
int myrand(void) {
next = next * 1103515245 + 12345;
return((unsigned)(next/65536) % 32768);
}
void mysrand(unsigned seed) {
next = seed;
}
It is trivially portable to Java. The quality of randomness produced this generator is not very high, but it's not really guaranteed to be any better with rand() either, so you would need to implement something fancier if better randomness is required.
In both C and Java, the same seed will generate the same random values. While the underlying mechanism might be different, this property is maintained in every programming language I know about.
C and Java will probably not generate the same set of random numbers given a seed. The only property that is maintained is that given a seed, a programming language will generate the same random numbers.
Related
I was reading java.util.Random class and noticed that there is no nextDouble(), nextFloat() and nextLong() which can accept a bound.
There are many way to get it done like this.
But my question is why java did not provide us with these required method like nextInt(int n) which accept the bound.
Is there any specific reason they did not provide these methods?
A good API always tries to provide the essential elements that a user needs to do his job.
Having nextInt(int n) is just one possible implementation. What if you need other distributions?!
In other words: the Random API could try to anticipate all potential usage patterns, but that would very much bloat the whole API. Instead, the designers choice a very small interface - but you got all the elements required to build your own things on top of that.
Thing is: in the end, this is a design style decision by the people who created the Random class. And as so often, problems could be solved in many different ways. Thus you shouldn't draw deep conclusions on the solution that was picked here.
Looking at the code from Random.java (in jdk 8) there are two statements that stand out.
* The algorithm is slightly tricky. It rejects values that would result
* in an uneven distribution (due to the fact that 2^31 is not divisible
* by n). The probability of a value being rejected depends on n.
and
* Linear congruential pseudo-random number generators such as the one
* implemented by this class are known to have short periods in the
* sequence of values of their low-order bits.
Without being an expert in random number generation (they refer to it as pseudo random number generation), it seems evident that the algorithm is trying to to a better job of returning "random" numbers than if you simply did next() % bound, in terms of randomness (and possibly also efficiency).
There is also the convenience factor but that doesn't seem to be the primary reason, given the comments in the code.
Whenever we create an object of Random class in java. We either of the constructor
Random()
Random(long seed)
What is the purpose of seed here in the 2nd constructor and how can I use it to my benefit i.e. manipulate its use?
The answer above sums it up clearly. As per java api docs from oracle, the first constructor
Random()
"Creates a new random number generator. This constructor sets the seed of the random number generator to a value very likely to be distinct from any other invocation of this constructor. "
The seed is probably a derivative of the current time, or the current time itself. That should be enough to be "very likely to be distinct from any other invocation". Which, in essence, is most likely what you need, most of the time.
So why have another constructor that takes a seed?
Simply put, if you want to generate the same set of random numbers over and over, you use the same seed on your Random constructor. This is useful when doing experiments on different control sets, and you don't want to bother creating your own table of random inputs, but still want the same set of random input on a different experiment/control set.
There's no such thing as truly random numbers in computing. The available methods for getting a random number across all programming languages is nothing but an algorithm to simulate random numbers.
In some languages (C++, I know for sure), an unseeded random number generator will return the same series of numbers on every fresh execution of the program.
What is common is to seed the random number generator with the current time (which will be random enough for most purposes) so that the algorithm starts with a random number each time.
Pseudo-random number generators maintain some set of state information, which is advanced through some recurrence relation to determine the next value of the state. The output of a PRNG is some function of the state. Java's Random class uses a Linear Congruential Generator. LCG's work using the recurrence relationship Ui+1 = (A Ui + C) % M for some constant integer values A, C, and M. Java's current implementation uses a 48-bit state but uses 32 bits or less of it on each iteration of the recurrence.
Based on this, you can see that if you start with the same state you will get the exact same sequence of values out of your PRNG. This can be useful if you want to be able to reproduce exactly the same sequence of "randomness", for instance for debugging or for comparing two experiments head-to-head.
If you invoke the constructor without an argument, it picks a starting value for the state with a promise that different invocations are very likely to be distinct from each other. If you supply a seed to the constructor, that seed's value is used to set the initial state.
If a seed number is defined for the random number generation, is it possible that different random number sequences are achieved on different computers? If so, how to achieve the same sequences?
private static final long seed = 1;
Random generator = new Random(seed);
for (int i = 0; i < nchrom; i++) {
val = (int) Math.round(generater.nextDouble()*(nchrom-1));
//...
}
Yes, with the same seed you should get the same sequence of numbers. The algorithm is specified in the documentation:
An instance of this class is used to generate a stream of pseudorandom numbers. The class uses a 48-bit seed, which is modified using a linear congruential formula. (See Donald Knuth, The Art of Computer Programming, Volume 2, Section 3.2.1.)
If two instances of Random are created with the same seed, and the same sequence of method calls is made for each, they will generate and return identical sequences of numbers. In order to guarantee this property, particular algorithms are specified for the class Random. Java implementations must use all the algorithms shown here for the class Random, for the sake of absolute portability of Java code. However, subclasses of class Random are permitted to use other algorithms, so long as they adhere to the general contracts for all the methods.
My only concern would be that if you're using nextDouble() you could run into some artifacts of floating point unit differences. I suspect you won't, but that would be my concern. I'd recommend that you use nextInt anyway:
val = generator.nextInt(nchrom); // Exclusive upper bound
Will two java.security.SecureRandom instances which are seeded with the same value initially give the same sequence of random numbers?
I am asking this because I want the same sequence of random numbers in both the client and the server. What if both of them are using the same seed value. Will the sequence be the same, or is there any way that the sequence can be made the same?
From the API docs:
If two instances of Random are created with the same seed, and the same sequence of method calls is made for each, they will generate and return identical sequences of numbers.
What if both of them are using the same seed value. Will the sequence be the same?
No, they definitely won't. At least not in Oracle's Java 7 SDK implementation. See my sample code in this SO post. It appears that the implementation may elect to use additional sources of randomness, in addition to the provided seed.
On Linux I am used to using mkpasswd to generate random passwords to use, on OS X however I don't have this command. Instead of sshing in to my vps every time, I wanted to re implement it using Java. What I have done is pick at random 4 lower case letters, 2 upper case letters, 2 symbols (/ . , etc) and 2 numbers. Then I create a vector and shuffle that too.
Do you think this is good enough randomization?
If you use java.security.SecureRandom instead of java.util.Random then it's probably secure. SecureRandom provides a "cryptographically strong pseudo-random number generator (PRNG)". I.e. it ensures that the seed cannot easily be guessed and that the numbers generated have high entropy.
yes, it is. If you are using java.util.Random:
An instance of this class is used to generate a stream of pseudorandom numbers. The class uses a 48-bit seed, which is modified using a linear congruential formula. (See Donald Knuth, The Art of Computer Programming, Volume 2, Section 3.2.1.)
The algorithms implemented by class Random use a protected utility method that on each invocation can supply up to 32 pseudorandomly generated bits.
EDIT
in response to a comment:
/**
* Creates a new random number generator. This constructor sets
* the seed of the random number generator to a value very likely
* to be distinct from any other invocation of this constructor.
*/
public Random() {
this(++seedUniquifier + System.nanoTime());
}
private static volatile long seedUniquifier = 8682522807148012L;
There is a similar pwgen command available in the Mac Ports.
Depends on where your entropy comes from. Using rand() or similar functions that your particular language comes with may not be secure.
On OSX you can use /dev/random I think.
It might be OK, but you should allow some randomization in password lengths perhaps.
If your program became popular it would become a weakness that the password length was public knowledge. Also randomize the exact ratio of lowercase:uppercase:symbols:numbers a little.
Why not just compile mkpasswd on your OS X host?