Home | Contact Us | FAQ | Search & Site Map | Link to Us
Sign In | Join | Other 45 Sites in Network
HomeAnnouncementsWhite Papers
Discussion GroupsFirst AidDatabasesJavaBeansGUIJava 3DVirtual MachineCORBASecurityToolsGeneral
Java DirectoryOpen Source ProjectsSample Book ChaptersUser GroupsWeb Resources
Related Topics
Databases.NETMore Topics ...

Java Forum / General / May 2006

Tip: Looking for answers? Try searching our database.

how to use recursive algorithm to determine all of the arrangements?

Thread view: 
index - 03 May 2006 02:55 GMT
suppose there are n elements,how can you use recursive algorithm to
determine all of the n elements' arrangements.
for example,if there is 1 element,say, A1,there is only one
arrangement, namely,{A1};
                 if there are 2 elements,say,A1 and A2,there are two
arrangements,namely, {A1,A2},{A2,A1};
   3 elements, A1,A2 and A3, six arrangements
exist,{A1,A2,A3},{A2,A1,A3},{A3,A1,A2},{A3,A2,A1},{A1,A3,A2},{A2,A3,A1};
.......
so if there are n elements, how can we use recursive algorithm to
determine all of these arrangements?
index - 03 May 2006 03:12 GMT
of course We know that there are n! permutations of n elements,but i
just want to make a program to print what does every permutation look
like,like
this,{A1,A2,A3},{A2,A1,A3},{A3,A1,A2},{A3,A2,A1},{A1,A3,A2},{A2,A3,A1}.
can anyone give me some tips?
James McGill - 03 May 2006 04:54 GMT
I'm stumped on this one.  I draw a circular figure in my panel, with the
midpoint of the circle being (radius, radius).  Then I want to draw
another circle over it in another layered pane, but rotated with respect
to the midpoint of the circles, not with respect to the panel.

I've tried lots and lots of combinations of translate and rotate, but I
just can't figure out what to do to make this happen.  How do you rotate
a circle with respect to its midpoint?
Rhino - 03 May 2006 05:25 GMT
> I'm stumped on this one.  I draw a circular figure in my panel, with the
> midpoint of the circle being (radius, radius).  Then I want to draw
[quoted text clipped - 4 lines]
> just can't figure out what to do to make this happen.  How do you rotate
> a circle with respect to its midpoint?

Uhh, how do you tell if it worked? If I draw a circle of radius 5 units
centered on (0,0), then rotate the circle 90 degrees, how does the second
circle look any different that the first?

If this is a homework assignment, maybe you've read it incorrectly??

--
Rhino
James McGill - 04 May 2006 01:08 GMT
> > I've tried lots and lots of combinations of translate and rotate, but I
> > just can't figure out what to do to make this happen.  How do you rotate
[quoted text clipped - 3 lines]
> centered on (0,0), then rotate the circle 90 degrees, how does the second
> circle look any different that the first?

It's a circular image with a pattern, not just a circle.  I was trying
to keep my question simple and focused on how to do rotates.

> If this is a homework assignment, maybe you've read it incorrectly??

I'm writing a game and I'm fairly new with the Java2D api.  I'm just now
figuring out how to deal with JLayeredPanes.   So one of the elements of
my game is an image on one layer with another moving image on its own
layer.  

Thanks,

James
Simon - 03 May 2006 08:59 GMT
> I'm stumped on this one.  I draw a circular figure in my panel, with the
> midpoint of the circle being (radius, radius).  Then I want to draw
> another circle over it in another layered pane, but rotated with respect
> to the midpoint of the circles, not with respect to the panel.

I'm not sure about your question, so let me restate it as I understand it. I
assume that you do not want to rotate the circle itself (because that doesn't
make much sense) but rather the center of a second object (which is again a
circle?) with respect to the origin of the first.

You have a circle at point (r,r) with radius r, i.e., it looks like this:
(use a fixed width font)

|           o
|  ________/
| /       / \
|/       /   \
|       /     |
r|      x      |
|             |
|\           /
|_\_________/____
       r

Now you want to draw another figure that is positioned relative to the center of
the circle at position o given in terms of an angle and a distance d from the
center x (perhaps you want it on the circumference, i.e., d=r)? Then you can do
the following:

1. Translate the graphics context to the center of the circle, i.e. translate(r,r)
2. Perform your rotation (around the origin which now coincides with the center
of the circle)
3. Translate the origin again to move away from the origin by d arriving at "o"
in the above drawing. (E.g., call translate(d,0).)
4. Draw the second object.

Is that what you want to do?

> I've tried lots and lots of combinations of translate and rotate, but I
> just can't figure out what to do to make this happen.  How do you rotate
> a circle with respect to its midpoint?
 ^^^^^^^^
Just to be sure: Maybe your problems are caused by a misunderstaning in the
usage of Graphics: calling translate() or rotate() or any other method of
Graphics does not move or rotate the things you have already drawn. You cannot
translate or rotate the circle but rather the graphics context itself. This has
an influence only on the objects you draw afterwards.

Hope this helps,
Simon
James McGill - 03 May 2006 18:01 GMT
> > I'm stumped on this one.  I draw a circular figure in my panel, with the
> > midpoint of the circle being (radius, radius).  Then I want to draw
[quoted text clipped - 5 lines]
> make much sense) but rather the center of a second object (which is again a
> circle?) with respect to the origin of the first.

Thanks for your answer, which already has helped, but let me clarify:

I make a buffered image for the base layer, which is a circle.  
Then I make another buffered image for the top layer, which is also a
circle, the same size.  

I want to animate the second image on top of the first, I guess I'd call
it a "pinwheel" effect.  So each timer tick, the top layer should be
redrawn in its next step of rotation.  
This worked for my purpose:
   
    ...
    // first translate
    g2.translate(radius, radius);
    // then rotate
    g2.rotate(stepRadians);
    // then translate back
    g2.translate(-radius, -radius);
    g2.drawImage(topLayerImage, 0, 0, this);
    ...

So I think I have to translate the origin back after rotating becuase
the image itself is oriented to that space, correct?
Simon - 04 May 2006 08:38 GMT
> I make a buffered image for the base layer, which is a circle.  
> Then I make another buffered image for the top layer, which is also a
[quoted text clipped - 17 lines]
> So I think I have to translate the origin back after rotating becuase
> the image itself is oriented to that space, correct?

Ok, now I see the problem. You are drawing an image and that does not have the
center at the "origin" but at the upper left corner. I thought you were drawing
a Shape which typically has the center at the origin. Therefore, you need the
second translation. I would assume that your code should actually solve the
problem, doesn't it?

One suggestion: Once you are sure about which transformations you need,
aggregate all of them into a single AffineTransform and apply it once when
rendering the image. There are also variants of the drawImage methods that
accept an AffineTransform. Depending on how quickly your wheel rotates and how
much memory you can affort, it might be even better to generate rotated copies
of the image in a preprocessing phase so you don't have to perform the time
consuming rotation for each frame.

Cheers,
Simon
Thomas Weidenfeller - 04 May 2006 09:03 GMT
> I make a buffered image for the base layer, which is a circle.  
> Then I make another buffered image for the top layer, which is also a
> circle, the same size.  
>
> I want to animate the second image on top of the first,

I would pre-calculate a set of rotated images. The number of
pre-calculated images should depend on the desired animation speed, and
the fps rate with which your game works.

Once you have them pre-calcualated you just have to select the one
matching the the current desired wheel position and draw it with a fixed
 x/y offset.

As a general hint, you are suggesting you are using JLayeredPane to
construct your animation. This doesn't sound too good. Java games
typically use multi-buffers, all manged by the game, and flip the
buffers manually. In your case one would probably compose each frame
from a fixed background buffer, plus the current animation and then flip
the buffer.

/Thomas
Signature

The comp.lang.java.gui FAQ:
ftp://ftp.cs.uu.nl/pub/NEWS.ANSWERS/computer-lang/java/gui/faq
http://www.uni-giessen.de/faq/archiv/computer-lang.java.gui.faq/

James McGill - 04 May 2006 17:42 GMT
> As a general hint, you are suggesting you are using JLayeredPane to
> construct your animation. This doesn't sound too good.  

It's not.  I need to move away from Swing/AWT entirely, I think.  

Thanks,

James.
Simon - 03 May 2006 08:32 GMT
> of course We know that there are n! permutations of n elements,but i
> just want to make a program to print what does every permutation look
> like,like
> this,{A1,A2,A3},{A2,A1,A3},{A3,A1,A2},{A3,A2,A1},{A1,A3,A2},{A2,A3,A1}.
>  can anyone give me some tips?

Remove the last element An and consider all arrangements for the reduced set.
Then insert An at every possible position.

Cheers,
Simon
Oliver Wong - 08 May 2006 22:24 GMT
> of course We know that there are n! permutations of n elements,but i
> just want to make a program to print what does every permutation look
> like,like
> this,{A1,A2,A3},{A2,A1,A3},{A3,A1,A2},{A3,A2,A1},{A1,A3,A2},{A2,A3,A1}.
> can anyone give me some tips?

   How would you, as a human being, solve this problem? I.e. how did you
come up with the above list of 6 elements?

   - Oliver
opalpa@gmail.com opalinski from opalpaweb - 09 May 2006 06:45 GMT
The following code is not recursive.  There are three files involved:

IndexMapper.java  Permutations.java  PossibleSets.java

btw, How do I check for overflow on multiplication in Java (or any
other arithmetic operation)?

One aspect I like about this code is that returned items are of same
type as passed in items, that is one can give an array of integers and
get arrays of integers, or one can give strings and get back strings.

Here is IndexMapper:
import java.util.*;
import java.lang.reflect.Array;
class IndexMapper implements Iterator {
 private Object map[];
 private Iterator it;
 private Class type = null;
 IndexMapper(Object map[], Iterator indicies) {
   this.map = map;
   this.it = indicies;
   type = map.getClass().getComponentType();
 }
 public Object next() {
   int current[] = (int []) it.next();
   Object trans[] = null;
   trans = (Object[]) Array.newInstance(type,current.length);

   for (int i=0; i<trans.length; i++)
     trans[i] = map[current[i]];
   return trans;
 }
 public boolean hasNext() {
   return it.hasNext();
 }
 public void remove() {
   it.remove();
 }
}

Here is PossibleSets:
import java.util.*;
/**
 This class enumerates subsets of arbitrary sets;
 The iterator operations are not synchronized.
*/
final public class PossibleSets {
 private int n, k;
 private Object map[];
 public PossibleSets(int from, int choose) {
   n = from;
   k = choose;
   if (n<=0 || k<=0)
     throw new IllegalArgumentException(
     "Cannot generate sets from zero or negative parameters.");
   if (k>n)
     throw new IllegalArgumentException(
     "Too many elements asked for.");
   map = null;
 }
 public PossibleSets(Object from[], int choose) {
   this(from.length,choose);
   map = from;
 }

 public long count() {
   long c = 1;
   long local_n = n;
   long local_k = k;

   while (local_k > 0)
   {
       c = c * local_n;
       if (c < 0)
         throw new IllegalArgumentException("Number too large.");
       --local_n;
       --local_k;
   }
   local_k = k;
   while (local_k > 0)
   {
       c = c / local_k;
       --local_k;
   }

   return c;
 }

 public Iterator iterator() {
   if (map == null)
     return new Provider(n,k);
   else
     return new IndexMapper(map, new Provider(n,k));
 }
 private class Provider implements Iterator {
   private int set[];
   private boolean primedWithAValue = false;
   /** @return int [] */
   public Object next() {
     if (! primedWithAValue)
       throw new NoSuchElementException("No more sets to access.");
     int returnSet[] = new int[k];
     for (int i=0;i<k;i++)
       returnSet[i] = set[i];
     prepareForCalls();
     return returnSet;
   }
   public boolean hasNext() {
     return primedWithAValue;
   }
   Provider(int n, int k) {
     set = new int[k];
     for (int i=0; i<k; i++) {
       set[i] = i;
     }
     primedWithAValue = true;
   }

   private void prepareForCalls() {
     if ( thereAreMorePossibleSets() )
       {
         primedWithAValue = true;
       }
     else
       {
         primedWithAValue = false;
       }
   }
   private boolean thereAreMorePossibleSets() {
       // generate next combination in lexicographical order
       int i = k - 1;                      // start at last item

       while (i>=0 && set[i] == (n - k + i))  // find next item to
increment
           --i;

       if (i < 0) return false;
       ++set[i];

       for (int j = i + 1; j < k; j++)
           set[j] = set[i] + j - i;
       return true;
   }
   public void remove() { throw new UnsupportedOperationException(); }
 }
 static private ArrayList arrayOfIntegers(Object o) {
   int a[] = (int []) o;
   ArrayList al = new ArrayList();
   for (int i=0; i<a.length; i++)
     al.add(new Integer(a[i]));
   return al;
 }
 static public void main(String args[]) {
   PossibleSets s = new PossibleSets(new Integer(args[0]).intValue(),
new Integer(args[1]).intValue());
   System.out.println("setcounts="+s.count());
   for (Iterator it=s.iterator(); it.hasNext(); )
     System.out.println("next="+arrayOfIntegers(it.next()));
 }
}

Here is Permutations:
import java.util.*;
final public class Permutations {
 private int n, k;
 private Object map[];
 public Permutations(int from, int choose) {
   n = from;
   k = choose;
   if (n<=0 || k<=0)
     throw new IllegalArgumentException("Cannot generate permutations
from zero or negative parameters.");
   if (k>n)
     throw new IllegalArgumentException("Too many elements asked
for.");
   map = null;
 }
 public Permutations(Object o[], int choose) {
   this(o.length,choose);
   map = o;
 }
 public long count() {
   long c = 1;
   long local_n = n;
   long local_k = k;

   while (local_k > 0)
   {
      c = c * local_n;
      if (c<0)
        throw new IllegalArgumentException("Number too large.");
       --local_n;
       --local_k;
   }

   return c;
 }

 public Iterator iterator() {
   if (map == null)
     return new Provider(n,k);
   else
     return new IndexMapper(map, new Provider(n,k));
 }
 private class Provider implements Iterator {
   private int perm[];
   private boolean primedWithAValue = false;
   private Iterator sets = null;
   /** @return int [] */
   public Object next() {
     if (! primedWithAValue)
       throw new NoSuchElementException("No more permutations to
access.");
     int returnPerm[] = new int[k];
     for (int i=0; i<k; i++)
       returnPerm[i]=perm[i];
     prepareForCalls();
     return returnPerm;
   }
   public boolean hasNext() {
     return primedWithAValue;
   }
   Provider(int n, int k) {
     sets = new PossibleSets(n,k).iterator();
     perm =  (int []) sets.next();
     primedWithAValue = true;
   }

   private void prepareForCalls() {
     if ( thereAreMorePermutationsInCurrentSet() )
       {
         int i = k - 1;
         while (perm[i-1] >= perm[i])
         i--;
         int j = k;
         while (perm[j-1] <= perm[i-1])
         j--;

         swap(i - 1, j - 1);

         i++;
         j = k;

         while (i < j)
         {
            swap(i - 1, j - 1);
            i++;
            j--;
         }
         primedWithAValue = true;
       }
     else if (thereAreMoreSets())
       {
         perm = (int[]) sets.next();
         primedWithAValue = true;
       }
     else
       {
         primedWithAValue = false;
       }
   }
   private void swap(int a, int b)
   {
     int temp  = perm[a];
     perm[a] = perm[b];
     perm[b] = temp;
   }
   private boolean thereAreMorePermutationsInCurrentSet() {
     int i = k - 1;
     while (i>=1 && (perm[i-1] >= perm[i]))
       i--;

     return i >= 1;
   }
   private boolean thereAreMoreSets() {
     return sets.hasNext();
   }
   public void remove() { throw new UnsupportedOperationException(); }
 }

 static private ArrayList arrayOfIntegers(Object o) {
   int a[] = (int []) o;
   ArrayList al = new ArrayList();
   for (int i=0; i<a.length; i++)
     al.add(new Integer(a[i]));
   return al;
 }
 static public void main(String args[]) {
   Permutations p = new Permutations(new Integer(args[0]).intValue(),
new Integer(args[1]).intValue());
   System.out.println("permcounts="+p.count());
   for (Iterator it=p.iterator(); it.hasNext(); )
     System.out.println("next="+arrayOfIntegers(it.next()));
 }
}
opalpa@gmail.com opalinski from opalpaweb - 09 May 2006 06:46 GMT
Opalinski
opalpa@gmail.com
http://www.geocities.com/opalpaweb/
Chris Uppal - 09 May 2006 15:40 GMT
opalpa@gmail.com wrote:

> btw, How do I check for overflow on multiplication in Java (or any
> other arithmetic operation)?

You might find the book mentioned in the class JavaDoc for java.lang.Integer
intersting.

   -- chris
Oliver Wong - 09 May 2006 15:48 GMT
<opalpa@gmail.com> wrote in message
news:1147153506.499376.122360@e56g2000cwe.googlegroups.com...

> btw, How do I check for overflow on multiplication in Java (or any
> other arithmetic operation)?

   The simplest way is to check for integer overflow is to do the
multiplication in floating point precision, and check if the result is
larger than what would fit in whatever primitive type you're trying to store
the value in.

e.g.

<pseudocode>
int a, b, result;
double temp = (double)a * (double)b;
if (temp > max_int_value) {
 handleOverFlow();
} else {
 result = (int)temp;
}
</pseudocode>

In the case of int, you could use long instead of double.

   - Oliver
Patricia Shanahan - 09 May 2006 16:44 GMT
> <opalpa@gmail.com> wrote in message
> news:1147153506.499376.122360@e56g2000cwe.googlegroups.com...
[quoted text clipped - 22 lines]
>
>    - Oliver

Note that this should only be used, as written, for int and smaller
types, not long. It depends on all int values being exactly
representable as doubles, so that if there is no overflow temp does
contain the exact answer.

This program:

public class TestMult {
  public static void main(String[] args) {
    long a = 0x0fffffffffffffffL;
    long b = 3;
    long result=0;
    double temp = (double)a * (double)b;
    if (temp > Long.MAX_VALUE) {
     handleOverFlow();
     System.exit(3);
    } else {
     result = (long)temp;
    }
    System.out.println("result="+result+" correct answer="+a*b);
  }
  private static void handleOverFlow() {
    System.out.println("Overflow detected");
  }
}

prints:

result=3458764513820540928 correct answer=3458764513820540925

For long, using BigInteger instead of double would be slower but more
reliable.

Patricia
Chris Uppal - 09 May 2006 18:29 GMT
> Note that this should only be used, as written, for int and smaller
> types, not long. It depends on all int values being exactly
> representable as doubles, so that if there is no overflow temp does
> contain the exact answer.

I think you can get a lot closer while still staying with the same basic
approach:

=============
   private static final double
       fmax = (double)Long.MAX_VALUE,
       fmin = (double)Long.MIN_VALUE;

   public static final long
   multiply(long a, long b)
   throws OverflowException
   {
       double
           floatA = (double)a,
           floatB = (double)b,
           floatC = floatA * floatB;

       if (floatC < fmin
       ||  floatC > fmax)
           throw new OverflowException("" + floatC);

       return a * b;
   }

=============

I don't think that can ever fail to notice an overflow condition, and whenever
it does return a value that value is correct.  I think it can only get it wrong
(i.e. throw when it doesn't have to) are for values of floatC which are > fmax
due /only/ to rounding, and similarly at the other end of the range.  I'm not
sure how many such values there are (nor how many floating-point
multiplications can actually yield them) but an ULP is only 2048 at that range,
so I don't think there can be many.

An alternative formulation which is very marginally slower, but which I don't
think has the false-negative problem at all is:

=============
   public final long
   multiply(long a, long b)
   throws OverflowException
   {
       long c = a * b;
       double fTrue = (double)a * (double)b;
       double fMaybe = (double)c;

       if (fTrue != fMaybe)
           throw throwable;

       return c;
   }

=============

But I'm not really sure whether that is actually correct for all possible
multiplications...

   -- chris
opalpa@gmail.com opalinski from opalpaweb - 11 May 2006 13:23 GMT
Thank you all for your technique.

Opalinski
opalpa@gmail.com
http://www.geocities.com/opalpaweb/


Free Magazines

Get these publications absolutely FREE for up to 12 months. There are no hidden fees and no obligation. Simply choose a title, complete the application form and submit it. Read more ...

Oracle MagazineNetwork ComputingComputer WorldBio-IT WorldeWeekInformation WeekInfosecurity
 
Sign In
Join
My Latest Posts
My Monitored Threads
My Blog
My Photo Gallery
My Profile
My Homepage

Start New Thread
Enable EMail Alerts
Rate this Thread



©2009 Advenet LLC   Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.