http://www.codechef.com/FEB10/problems/M5/

On Wed, Jul 20, 2011 at 5:35 PM, Piyush Sinha <[email protected]>wrote:

> @Dumanshu..i am not partitioning them into just two queues...
>
> Moreover I just gave a raw idea...and yeah the complexity is in the order
> of n^2 only.....
> There are many chances of improvement in it..
>
> On Wed, Jul 20, 2011 at 5:30 PM, Dumanshu <[email protected]> wrote:
>
>> @Piyush:
>> Initially for partitioning the given circles into the 2 queues u r
>> having an O(n^2) loop, so u are comparing each circle with every
>> other.
>> Now, it is possible that u have 3 or more circles A,B,C intersecting
>> if i got ur algo correct, ur intersection queue will have AB, BC, CA.
>> So, according to the geometry, u will find the areas. But this area
>> would be different than the actual area for intersection of A,B,C.
>>
>> On Jul 20, 3:48 pm, Piyush Sinha <[email protected]> wrote:
>> > I would like to redefine my algo with cases clarified...
>> >
>> > Create a queue that is made to contain the points...
>> >
>> > say points queue [1000];
>> >
>> > for i:1 to n
>> >  for j:i+1 to n
>> >      Calculate d (distance between the two centers)
>> >      if (d >= r0 + r1) keep them in two separate queues //the circles
>> > don't intersect
>> >      if(d==0 || d<= abs(r0-r1))
>> >              ignore the circle with smaller radius // one circle
>> > wholly contains another such that  the borders do not overlap, or
>> > overlap exactly (e.g. two identical circles)
>> >      else
>> >           keep both of them in one single queue
>> >
>> > Now calculate the area of the circles in those queues which have
>> > single element...
>> >
>> > those with more than one element..calculate the area using simple
>> > geometry...You can take help of this..
>> http://mathworld.wolfram.com/Circle-CircleIntersection.html
>> >
>> > Hope its clear now...
>> >
>> > On 7/20/11, SAMMM <[email protected]> wrote:
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> >
>> > > I doubth .
>> >
>> > > For (d< r0 + r1) ignore the point with smaller radius as it will
>> > > overshadowed the bigger circle completely
>> >
>> > > There may be a case where the circle is partially overlapped by the
>> > > other circles. Then this algo will fail .
>> >
>> > > The area will be of like these :-
>> >
>> > > Suppose 3 circles are there X,Y&Z .
>> > > Then the area will be :-
>> >
>> > > Case1:-  X+Y+Z
>> > > Case2:-  X+(YUZ) ==>> Y + Z - (YnZ) <--- intersection
>> > > case3:- There circle can overlap ... like these .
>> >
>> > > Then Will your algo work .. I guess no .
>> >
>> > > --
>> > > You received this message because you are subscribed to the Google
>> Groups
>> > > "Algorithm Geeks" group.
>> > > To post to this group, send email to [email protected].
>> > > To unsubscribe from this group, send email to
>> > > [email protected].
>> > > For more options, visit this group at
>> > >http://groups.google.com/group/algogeeks?hl=en.
>> >
>> > --
>> > *Piyush Sinha*
>> > *IIIT, Allahabad*
>> > *+91-7483122727*
>> > * <https://www.facebook.com/profile.php?id=100000655377926> "NEVER SAY
>> > NEVER"
>> > *
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "Algorithm Geeks" group.
>> To post to this group, send email to [email protected].
>> To unsubscribe from this group, send email to
>> [email protected].
>> For more options, visit this group at
>> http://groups.google.com/group/algogeeks?hl=en.
>>
>>
>
>
> --
> *Piyush Sinha*
> *IIIT, Allahabad*
> *+91-7483122727*
> * "NEVER SAY NEVER"
> *
>
> --
> You received this message because you are subscribed to the Google Groups
> "Algorithm Geeks" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected].
> For more options, visit this group at
> http://groups.google.com/group/algogeeks?hl=en.
>

-- 
You received this message because you are subscribed to the Google Groups 
"Algorithm Geeks" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/algogeeks?hl=en.

Reply via email to