Solved

Really tricky algorithm question

Posted on 2004-04-05
18
354 Views
Last Modified: 2008-02-20
My job unfortunatly just got rid of some really talented programmers, and i am left to do this. We are working with a new system we are developing which takes an input (stock quotes) from a database (could even be excel file) and returns a specific answers based on the following function:
The function to be evaluated is this:  P.S. the 'E' means Summation (Sigma).

         N-1                    L-1         M-1
y(n) = E a(k) x(n-k) +     E     +     E  b(l,m) x(n-1)  x(n-m)
        K=0                     I=0        m=0


Now, i have no idea why the algorithm has to be based on the above function, (remember, i am taking over a half completed job and im not even a real programmer) but this is what was in the old programmers notes. I need some sort of java or c++ code to evaluate the above function. I know that we are going to be working with 200 stock quotes at a time, so let x(n) be 200 values, choosing some variable (a and b) to be real constant coefficiants.
0
Comment
Question by:enray
  • 7
  • 6
  • 3
  • +1
18 Comments
 
LVL 12

Accepted Solution

by:
venkateshwarr earned 125 total points
ID: 10762434
Assuming there are values in arrays

public float y(n)
{
t1=0;
for(int k=0;k<=N-1;k++)
{
   t1=t1+a[k]*x[n-k];
}

for(int l=0;l<=L-1;l++)
for(int m=0;l<=M-1;m++)
{
   t1=t1+ b[l][m]*x[n-1]*x[n-m];
}
}

System.out.println("y100="+y(100));

0
 
LVL 12

Expert Comment

by:venkateshwarr
ID: 10762440
I assumed there are already values loaded in arrays a,x,b
0
 
LVL 14

Expert Comment

by:Tommy Braas
ID: 10763297
What's the name of the formula?
0
Courses: Start Training Online With Pros, Today

Brush up on the basics or master the advanced techniques required to earn essential industry certifications, with Courses. Enroll in a course and start learning today. Training topics range from Android App Dev to the Xen Virtualization Platform.

 

Author Comment

by:enray
ID: 10767450
Im not sure what the name is, remember i am reading a programmers notes, but i looked online and it seems like part of an Finitie Impulse Response Filter.
0
 
LVL 14

Expert Comment

by:Tommy Braas
ID: 10767938
In that case, you would want to have a look at the following file:
http://www.dspguru.com/sw/lib/fir_algs_1-0.c

It is an implementation in C and the fir_basic() function is pretty straight forward to re-implement in Java, like so:

public class FIR {

/**   These functions use most or all of the following input parameters:
*
*       input    - the input sample data
*       ntaps    - the number of taps in the filter  
*       h[]      - the FIR coefficient array
*       z[]      - the FIR delay line array
*/

/****************************************************************************
* fir_basic: Does the basic FIR algorithm: store input sample, calculate      
* output sample, move delay line                                          
*****************************************************************************/
public static double firBasic(double input, int ntaps, doubleh[], double z[])      
{
    int ii;
    double accum;
   
    /* store input at the beginning of the delay line */
    z[0] = input;

    /* calc FIR */
    accum = 0;
    for (ii = 0; ii < ntaps; ii++) {
        accum += h[ii] * z[ii];
    }

    /* shift delay line */
    for (ii = ntaps - 2; ii >= 0; ii--) {
        z[ii + 1] = z[ii];
    }

    return accum;
}
}

You then call the method by: FIR.firBasic(startValue, numTaps, doubleArray1, doubleArray2);

Replace "startValue", "numTaps", "doubleArray1", "doubleArray2" with the appropriate variables.
0
 

Expert Comment

by:deuel47
ID: 10767973
Well here's an idea, treat each function as a function, in other words convert each function in a java function.

Random example not based on anything:
a(k)
 public float a(k)
{
   \\does some work
    returns n;
}

do that for each function, then make a function to handle the entire summation.

You probably want to research the function to understand what it does in order to code it
 
0
 
LVL 14

Expert Comment

by:Tommy Braas
ID: 10768430
Deuel47, please keep in mind that enray is a self-proclaimed non-programmer.
0
 

Expert Comment

by:deuel47
ID: 10769581
yeah your right orangehead....sorry about that
0
 
LVL 14

Expert Comment

by:Tommy Braas
ID: 10769676
no worries! :-)
0
 

Author Comment

by:enray
ID: 10788901
Thanks guys... let me try out these code examples, then i will reward points. Thank you all so much!
0
 

Author Comment

by:enray
ID: 10789337
yenkateshwarr, what does t1 refer to? Also, which is the array which will hold the initial values loaded from the external file??
0
 
LVL 14

Expert Comment

by:Tommy Braas
ID: 10789642
I doubt that yenkateshwarr's example implements the actual algorithm. The source file link that I provided contains implementations of the algorithm with a bunch of variations.
0
 
LVL 12

Expert Comment

by:venkateshwarr
ID: 10793312
t1 is just a temporary double/float value...
this value is returned after the final calcuation is done
0
 

Author Comment

by:enray
ID: 10834300
Orangeheaed911, what are the nTaps???  
0
 
LVL 14

Expert Comment

by:Tommy Braas
ID: 10834698
It's the number of applications of the filter. I.e. the size of the arrays.
0
 

Author Comment

by:enray
ID: 10853858
Allright guys, this is what i have: I think it will be good for testing purposes of a specific FIR, but i am now getting indexoutofbounds errors.  Please advise:

class fir2 {
static public float y(int n)
{
int a[] = new int[200];
int x[] = new int[3];
int b[][]=  new int[4][3];

for(int i=0; i<199; i++) { // load array
      a[i] = i+1;
}

for(int i = 0;i<2;i++){ // load array
      x[i]=i+1;
}

for (int i = 0; i < 4; i++){ // load 2d array
       for (int j = 0; j < 3; j++){
              b[i][j] = i*j;
    }
}      

int t1=0;

for(int k=0;k<=n-1;k++)

{
   t1=t1+a[k]*x[n-k];
}

for(int l=0; l <= l-1; l++)
for(int m=0;l<=m-1;m++)
{
   t1=t1+ b[l][m]*x[n-1]*x[n-m];
}
return n;
}
public static void main(String[] args) {
      y(2);
      System.out.println("y100="+y(100));
}
0
 

Author Comment

by:enray
ID: 10966090
Thank you all... points to venkateshwarr
0
 
LVL 14

Expert Comment

by:Tommy Braas
ID: 10966235
enray, there is such a thing as a point split.

Btw, the reason why you're getting an ArrayIndexOutOfBoundsException is the fact that the array's are defined with a fixed size whereas the input value of n is not bound. This makes the line;
   t1=t1+a[k]*x[n-k];
break when n is larger than the size of a. Also, x is only 3 in size and the index in the earlier referenced line will go from n to n-k, which in your code is from 100 to 0, which is way outside the bounds of that array.
0

Featured Post

Live: Real-Time Solutions, Start Here

Receive instant 1:1 support from technology experts, using our real-time conversation and whiteboard interface. Your first 5 minutes are always free.

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

Java contains several comparison operators (e.g., <, <=, >, >=, ==, !=) that allow you to compare primitive values. However, these operators cannot be used to compare the contents of objects. Interface Comparable is used to allow objects of a cl…
Basic understanding on "OO- Object Orientation" is needed for designing a logical solution to solve a problem. Basic OOAD is a prerequisite for a coder to ensure that they follow the basic design of OO. This would help developers to understand the b…
This tutorial explains how to use the VisualVM tool for the Java platform application. This video goes into detail on the Threads, Sampler, and Profiler tabs.
Viewers will learn how to properly install Eclipse with the necessary JDK, and will take a look at an introductory Java program. Download Eclipse installation zip file: Extract files from zip file: Download and install JDK 8: Open Eclipse and …

785 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question