Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

the Problem of Apollonius:Calculating (x,y,r) of eight circles given 3 circles

6 views
Skip to first unread message

BookWight

unread,
May 12, 2013, 3:37:25 PM5/12/13
to
Hi -

Is it possible to calculate the (x,y) co-ordinates and the radius of each
of the eight circles, given three circles of known position and radius?

original 3 circles:

(ax1,ay1,ar1)
(ax2,ay2,ar2)
(ax3,ay3,ar3)

solutions:

(sx1,sy1,sr1)
(sx2,sy2,sr2)
etc

does it help to assume all 3 circles are in the 1st quadrant?

William Elliot

unread,
May 12, 2013, 9:54:30 PM5/12/13
to
On Sun, 12 May 2013, BookWight wrote:

> Is it possible to calculate the (x,y) co-ordinates and the radius of
> each of the eight circles, given three circles of known position and
> radius?
>
No, it's not possible to know anything about the other five circles.

> original 3 circles:
>
> (ax1,ay1,ar1)
> (ax2,ay2,ar2)
> (ax3,ay3,ar3)
>
> solutions:
>
> (sx1,sy1,sr1)
> (sx2,sy2,sr2)
> etc
>
> does it help to assume all 3 circles are in the 1st quadrant?
>
No. What helps is proper punctuation - beginning sentences with
a capital letter.

Christopher J. Henrich

unread,
May 13, 2013, 2:53:23 PM5/13/13
to
In article <XnsA1BE806A3...@216.151.153.167>, BookWight
<Book...@publicspam.com> wrote:

> Hi -
>
> Is it possible to calculate the (x,y) co-ordinates and the radius of each
> of the eight circles, given three circles of known position and radius?
Yes, because the problem is solvable with ruler & compass. Various
books on "advanced geometry" or "college geometry" discuss this. Also,
see /What/ /Is/ /Msthematics/ by Courant and Robbins, recently reissued
with added comments by Ian Stewart.
>
> original 3 circles:
>
> (ax1,ay1,ar1)
> (ax2,ay2,ar2)
> (ax3,ay3,ar3)
>
> solutions:
>
> (sx1,sy1,sr1)
> (sx2,sy2,sr2)
> etc
>
> does it help to assume all 3 circles are in the 1st quadrant?
I don't think so.

--
Chris Henrich
http://www.mathinteract.com
God just doesn't fit inside a single religion.

Ken Pledger

unread,
May 13, 2013, 5:40:30 PM5/13/13
to
In article <130520131453237591%chen...@monmouth.com>,
"Christopher J. Henrich" <chen...@monmouth.com> wrote:

> In article <XnsA1BE806A3...@216.151.153.167>, BookWight
> <Book...@publicspam.com> wrote:
> ....
> > Is it possible to calculate the (x,y) co-ordinates and the radius of each
> > of the eight circles, given three circles of known position and radius?
> Yes, because the problem is solvable with ruler & compass. Various
> books on "advanced geometry" or "college geometry" discuss this. Also,
> see /What/ /Is/ /Msthematics/ by Courant and Robbins, recently reissued
> with added comments by Ian Stewart....


The problem goes back to Apollonius. Courant & Robbins give an
elegant treatment by inversion, but their earlier discussion using
coordinates just explains why it can be done in principle, without
actually doing it.

There's an interestng little passage about this in E.T. Bell, "Men of
Mathematics", vol.1, p.52, about Descartes and his pupil Princess
Elisabeth. On the Apollonius circle problem, Bell says "This problem is
a fine specimen of the sort that are _not_ adapted to the crude brute
force of elementary Cartesian geometry. _Elisabeth solved it by
Descartes' methods._ It was rather cruel of him to let her do it...."

Ken Pledger.

Christopher J. Henrich

unread,
May 13, 2013, 7:20:31 PM5/13/13
to
In article
<ken.pledger-D344...@news.eternal-september.org>, Ken
Hmm, yes. But with the machinery of "Lie sphere geometry" (Wikipedia),
it becomes simple.
0 new messages