8 replies to this topic

[Dethlehem]

Hi, I have been trying to complete the following task for a while now:

"Every mathematician knows that the sum of the series
1 + 1/2 + 1/3 + 1/4 + ...
is infinite. But computers are funny things. Write a program which adds up the numbers in this series until the result of adding another one doesn’t change the sum. Use float variables for the arithmetic. Also, find out how many terms in the series are used.
[ The answer is 15.403683 and the number of terms added is 2097152. ] "



I have got this far:

class Terms {


    public static void main(String[] args) {


        int n=1;

        float lastFx = 0;

        float fx=1;


        while (lastFx != fx) {

            n++;

            lastFx = 1/n;

            fx = 1/(n+1);

        }


        System.out.println("number of terms used: " + n);

    }

}

from this, all the system prints out is: "number of terms used: 2" or "number of terms is 3", or 0.

Why can't i get this program working? :/

Any help is much appreciated

[WingedPanther]

You're doing integer division, not floating point division.

[Dethlehem]

WingedPanther said:

You're doing integer division, not floating point division.

ok thanks :), I corrected the mistake, all variables, including n are now of float type.

But now i get "number of terms used: 1.1864338E7", and not "2097152", and apparently, that's what it should be? o_O

[lethalwire]

You're losing some precision.
If you want to retain this precision use the BigDecimal class.

[wim DC]

If you'd use BigDecmal, the while loop lasts forever I think.

2 errors you have:

• Calculation of lastFx
  
• calculation of fx

lastFx is simple, just do

lastFx = fx;

I'll let you think again for the calculation of fx itself :P

Also note, that to declare a floating point number you must add the letter 'f' behind the number:

1 = int

1.0 = double

[B]1f = float

1.0f = float[/B]

1L = long

It's not required for

float fx = 1;

That will get autocasted, but in the formula of calculating fx you propably need it.

And finally: variable 'n' CAN be an integer. It'll get autocasted anyway.

And finally finally: 2097152 is indeed the correct answer

[Dethlehem]

wim DC said:

If you'd use BigDecmal, the while loop lasts forever I think.

2 errors you have:

• Calculation of lastFx
  
• calculation of fx

lastFx is simple, just do

lastFx = fx;

I'll let you think again for the calculation of fx itself :P

Also note, that to declare a floating point number you must add the letter 'f' behind the number:

1 = int

1.0 = double

[B]1f = float

1.0f = float[/B]

1L = long

It's not required for

float fx = 1;

That will get autocasted, but in the formula of calculating fx you propably need it.

And finally: variable 'n' CAN be an integer. It'll get autocasted anyway.

And finally finally: 2097152 is indeed the correct answer

Thanks for the reply, it has been the most helpful so far, but I still can't do this :/ I convinced myself it's my lack of understanding of float variables, because the series is easy as hell to understand o_O

this is how far i got now:

class Terms {


    public static void main(String[] args) {


        int n = 1;

        float lastFx = 0f;

        float fx=1f/n;


        while (lastFx != fx) {

        

            lastFx = fx;

        fx = 1f/n;

        n++;

        }


        System.out.println("number of terms used: " + n);

    }

}

& the output i get from this is "numer of terms used: 2" :S

[WingedPanther]

Step through your logic:
You set fx = lf/n
You set lastFx = fx
You set fx = lf/n (no change)
You increment n to 2
loop exits

[wim DC]

You still calculate fx wrong. It's not 1/n

[Dethlehem]

can someone give an answer to end my little bit of frustration with this question please? Because I just don't understand this :/