001/* Random.java -- a pseudo-random number generator
002   Copyright (C) 1998, 1999, 2000, 2001, 2002 Free Software Foundation, Inc.
003
004This file is part of GNU Classpath.
005
006GNU Classpath is free software; you can redistribute it and/or modify
007it under the terms of the GNU General Public License as published by
008the Free Software Foundation; either version 2, or (at your option)
009any later version.
010
011GNU Classpath is distributed in the hope that it will be useful, but
012WITHOUT ANY WARRANTY; without even the implied warranty of
013MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
014General Public License for more details.
015
016You should have received a copy of the GNU General Public License
017along with GNU Classpath; see the file COPYING.  If not, write to the
018Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
01902110-1301 USA.
020
021Linking this library statically or dynamically with other modules is
022making a combined work based on this library.  Thus, the terms and
023conditions of the GNU General Public License cover the whole
024combination.
025
026As a special exception, the copyright holders of this library give you
027permission to link this library with independent modules to produce an
028executable, regardless of the license terms of these independent
029modules, and to copy and distribute the resulting executable under
030terms of your choice, provided that you also meet, for each linked
031independent module, the terms and conditions of the license of that
032module.  An independent module is a module which is not derived from
033or based on this library.  If you modify this library, you may extend
034this exception to your version of the library, but you are not
035obligated to do so.  If you do not wish to do so, delete this
036exception statement from your version. */
037
038
039package java.util;
040
041import java.io.Serializable;
042
043/**
044 * This class generates pseudorandom numbers.  It uses the same
045 * algorithm as the original JDK-class, so that your programs behave
046 * exactly the same way, if started with the same seed.
047 *
048 * The algorithm is described in <em>The Art of Computer Programming,
049 * Volume 2</em> by Donald Knuth in Section 3.2.1.  It is a 48-bit seed,
050 * linear congruential formula.
051 *
052 * If two instances of this class are created with the same seed and
053 * the same calls to these classes are made, they behave exactly the
054 * same way.  This should be even true for foreign implementations
055 * (like this), so every port must use the same algorithm as described
056 * here.
057 *
058 * If you want to implement your own pseudorandom algorithm, you
059 * should extend this class and overload the <code>next()</code> and
060 * <code>setSeed(long)</code> method.  In that case the above
061 * paragraph doesn't apply to you.
062 *
063 * This class shouldn't be used for security sensitive purposes (like
064 * generating passwords or encryption keys.  See <code>SecureRandom</code>
065 * in package <code>java.security</code> for this purpose.
066 *
067 * For simple random doubles between 0.0 and 1.0, you may consider using
068 * Math.random instead.
069 *
070 * @see java.security.SecureRandom
071 * @see Math#random()
072 * @author Jochen Hoenicke
073 * @author Eric Blake (ebb9@email.byu.edu)
074 * @status updated to 1.4
075 */
076public class Random implements Serializable
077{
078  /**
079   * True if the next nextGaussian is available.  This is used by
080   * nextGaussian, which generates two gaussian numbers by one call,
081   * and returns the second on the second call.
082   *
083   * @serial whether nextNextGaussian is available
084   * @see #nextGaussian()
085   * @see #nextNextGaussian
086   */
087  private boolean haveNextNextGaussian;
088
089  /**
090   * The next nextGaussian, when available.  This is used by nextGaussian,
091   * which generates two gaussian numbers by one call, and returns the
092   * second on the second call.
093   *
094   * @serial the second gaussian of a pair
095   * @see #nextGaussian()
096   * @see #haveNextNextGaussian
097   */
098  private double nextNextGaussian;
099
100  /**
101   * The seed.  This is the number set by setSeed and which is used
102   * in next.
103   *
104   * @serial the internal state of this generator
105   * @see #next(int)
106   */
107  private long seed;
108
109  /**
110   * Compatible with JDK 1.0+.
111   */
112  private static final long serialVersionUID = 3905348978240129619L;
113
114  /**
115   * Creates a new pseudorandom number generator.  The seed is initialized
116   * to the current time, as if by
117   * <code>setSeed(System.currentTimeMillis());</code>.
118   *
119   * @see System#currentTimeMillis()
120   */
121  public Random()
122  {
123    this(System.currentTimeMillis());
124  }
125
126  /**
127   * Creates a new pseudorandom number generator, starting with the
128   * specified seed, using <code>setSeed(seed);</code>.
129   *
130   * @param seed the initial seed
131   */
132  public Random(long seed)
133  {
134    setSeed(seed);
135  }
136
137  /**
138   * Sets the seed for this pseudorandom number generator.  As described
139   * above, two instances of the same random class, starting with the
140   * same seed, should produce the same results, if the same methods
141   * are called.  The implementation for java.util.Random is:
142   *
143<pre>public synchronized void setSeed(long seed)
144{
145  this.seed = (seed ^ 0x5DEECE66DL) & ((1L &lt;&lt; 48) - 1);
146  haveNextNextGaussian = false;
147}</pre>
148   *
149   * @param seed the new seed
150   */
151  public synchronized void setSeed(long seed)
152  {
153    this.seed = (seed ^ 0x5DEECE66DL) & ((1L << 48) - 1);
154    haveNextNextGaussian = false;
155  }
156
157  /**
158   * Generates the next pseudorandom number.  This returns
159   * an int value whose <code>bits</code> low order bits are
160   * independent chosen random bits (0 and 1 are equally likely).
161   * The implementation for java.util.Random is:
162   *
163<pre>protected synchronized int next(int bits)
164{
165  seed = (seed * 0x5DEECE66DL + 0xBL) & ((1L &lt;&lt; 48) - 1);
166  return (int) (seed &gt;&gt;&gt; (48 - bits));
167}</pre>
168   *
169   * @param bits the number of random bits to generate, in the range 1..32
170   * @return the next pseudorandom value
171   * @since 1.1
172   */
173  protected synchronized int next(int bits)
174  {
175    seed = (seed * 0x5DEECE66DL + 0xBL) & ((1L << 48) - 1);
176    return (int) (seed >>> (48 - bits));
177  }
178
179  /**
180   * Fills an array of bytes with random numbers.  All possible values
181   * are (approximately) equally likely.
182   * The JDK documentation gives no implementation, but it seems to be:
183   *
184<pre>public void nextBytes(byte[] bytes)
185{
186  for (int i = 0; i &lt; bytes.length; i += 4)
187  {
188    int random = next(32);
189    for (int j = 0; i + j &lt; bytes.length && j &lt; 4; j++)
190    {
191      bytes[i+j] = (byte) (random & 0xff)
192      random &gt;&gt;= 8;
193    }
194  }
195}</pre>
196   *
197   * @param bytes the byte array that should be filled
198   * @throws NullPointerException if bytes is null
199   * @since 1.1
200   */
201  public void nextBytes(byte[] bytes)
202  {
203    int random;
204    // Do a little bit unrolling of the above algorithm.
205    int max = bytes.length & ~0x3;
206    for (int i = 0; i < max; i += 4)
207      {
208        random = next(32);
209        bytes[i] = (byte) random;
210        bytes[i + 1] = (byte) (random >> 8);
211        bytes[i + 2] = (byte) (random >> 16);
212        bytes[i + 3] = (byte) (random >> 24);
213      }
214    if (max < bytes.length)
215      {
216        random = next(32);
217        for (int j = max; j < bytes.length; j++)
218          {
219            bytes[j] = (byte) random;
220            random >>= 8;
221          }
222      }
223  }
224
225  /**
226   * Generates the next pseudorandom number.  This returns
227   * an int value whose 32 bits are independent chosen random bits
228   * (0 and 1 are equally likely).  The implementation for
229   * java.util.Random is:
230   * 
231<pre>public int nextInt()
232{
233  return next(32);
234}</pre>
235   *
236   * @return the next pseudorandom value
237   */
238  public int nextInt()
239  {
240    return next(32);
241  }
242
243  /**
244   * Generates the next pseudorandom number.  This returns
245   * a value between 0(inclusive) and <code>n</code>(exclusive), and
246   * each value has the same likelihodd (1/<code>n</code>).
247   * (0 and 1 are equally likely).  The implementation for
248   * java.util.Random is:
249   * 
250<pre>
251public int nextInt(int n)
252{
253  if (n &lt;= 0)
254    throw new IllegalArgumentException("n must be positive");
255
256  if ((n & -n) == n)  // i.e., n is a power of 2
257    return (int)((n * (long) next(31)) &gt;&gt; 31);
258
259  int bits, val;
260  do
261  {
262    bits = next(31);
263    val = bits % n;
264  }
265  while(bits - val + (n-1) &lt; 0);
266
267  return val;
268}</pre>
269   *   
270   * <p>This algorithm would return every value with exactly the same
271   * probability, if the next()-method would be a perfect random number
272   * generator.
273   *
274   * The loop at the bottom only accepts a value, if the random
275   * number was between 0 and the highest number less then 1<<31,
276   * which is divisible by n.  The probability for this is high for small
277   * n, and the worst case is 1/2 (for n=(1<<30)+1).
278   *
279   * The special treatment for n = power of 2, selects the high bits of
280   * the random number (the loop at the bottom would select the low order
281   * bits).  This is done, because the low order bits of linear congruential
282   * number generators (like the one used in this class) are known to be
283   * ``less random'' than the high order bits.
284   *
285   * @param n the upper bound
286   * @throws IllegalArgumentException if the given upper bound is negative
287   * @return the next pseudorandom value
288   * @since 1.2
289   */
290  public int nextInt(int n)
291  {
292    if (n <= 0)
293      throw new IllegalArgumentException("n must be positive");
294    if ((n & -n) == n) // i.e., n is a power of 2
295      return (int) ((n * (long) next(31)) >> 31);
296    int bits, val;
297    do
298      {
299        bits = next(31);
300        val = bits % n;
301      }
302    while (bits - val + (n - 1) < 0);
303    return val;
304  }
305
306  /**
307   * Generates the next pseudorandom long number.  All bits of this
308   * long are independently chosen and 0 and 1 have equal likelihood.
309   * The implementation for java.util.Random is:
310   *
311<pre>public long nextLong()
312{
313  return ((long) next(32) &lt;&lt; 32) + next(32);
314}</pre>
315   *
316   * @return the next pseudorandom value
317   */
318  public long nextLong()
319  {
320    return ((long) next(32) << 32) + next(32);
321  }
322
323  /**
324   * Generates the next pseudorandom boolean.  True and false have
325   * the same probability.  The implementation is:
326   * 
327<pre>public boolean nextBoolean()
328{
329  return next(1) != 0;
330}</pre>
331   *
332   * @return the next pseudorandom boolean
333   * @since 1.2
334   */
335  public boolean nextBoolean()
336  {
337    return next(1) != 0;
338  }
339
340  /**
341   * Generates the next pseudorandom float uniformly distributed
342   * between 0.0f (inclusive) and 1.0f (exclusive).  The
343   * implementation is as follows.
344   * 
345<pre>public float nextFloat()
346{
347  return next(24) / ((float)(1 &lt;&lt; 24));
348}</pre>
349   *
350   * @return the next pseudorandom float
351   */
352  public float nextFloat()
353  {
354    return next(24) / (float) (1 << 24);
355  }
356
357  /**
358   * Generates the next pseudorandom double uniformly distributed
359   * between 0.0 (inclusive) and 1.0 (exclusive).  The
360   * implementation is as follows.
361   *
362<pre>public double nextDouble()
363{
364  return (((long) next(26) &lt;&lt; 27) + next(27)) / (double)(1L &lt;&lt; 53);
365}</pre>
366   *
367   * @return the next pseudorandom double
368   */
369  public double nextDouble()
370  {
371    return (((long) next(26) << 27) + next(27)) / (double) (1L << 53);
372  }
373
374  /**
375   * Generates the next pseudorandom, Gaussian (normally) distributed
376   * double value, with mean 0.0 and standard deviation 1.0.
377   * The algorithm is as follows.
378   * 
379<pre>public synchronized double nextGaussian()
380{
381  if (haveNextNextGaussian)
382  {
383    haveNextNextGaussian = false;
384    return nextNextGaussian;
385  }
386  else
387  {
388    double v1, v2, s;
389    do
390    {
391      v1 = 2 * nextDouble() - 1; // between -1.0 and 1.0
392      v2 = 2 * nextDouble() - 1; // between -1.0 and 1.0
393      s = v1 * v1 + v2 * v2;
394    }
395    while (s >= 1);
396
397    double norm = Math.sqrt(-2 * Math.log(s) / s);
398    nextNextGaussian = v2 * norm;
399    haveNextNextGaussian = true;
400    return v1 * norm;
401  }
402}</pre>
403   *
404   * <p>This is described in section 3.4.1 of <em>The Art of Computer
405   * Programming, Volume 2</em> by Donald Knuth.
406   *
407   * @return the next pseudorandom Gaussian distributed double
408   */
409  public synchronized double nextGaussian()
410  {
411    if (haveNextNextGaussian)
412      {
413        haveNextNextGaussian = false;
414        return nextNextGaussian;
415      }
416    double v1, v2, s;
417    do
418      {
419        v1 = 2 * nextDouble() - 1; // Between -1.0 and 1.0.
420        v2 = 2 * nextDouble() - 1; // Between -1.0 and 1.0.
421        s = v1 * v1 + v2 * v2;
422      }
423    while (s >= 1);
424    double norm = Math.sqrt(-2 * Math.log(s) / s);
425    nextNextGaussian = v2 * norm;
426    haveNextNextGaussian = true;
427    return v1 * norm;
428  }
429}