Re: Packing problem
- From: "news.skynet.be" <bob@bob>
- Date: Tue, 31 Jan 2006 20:57:18 +0100
"news.skynet.be" <bob@bob> wrote in message news:43df9d0a$0$29466
> Given a finite set of infinite, one-dimensional bins and a collection of
> one-dimensional items. I want to distribute the items over the bins so
> that the variance of the bin contents is minimal. What algorithm do I use?
> (As a silly heuristic, I come up with: sort the items by length and put
> them, in order, from long to short, in the least filled bin)
I'll check out Robert's post.
Let me note that the above proposed wild guess is not correct.
Example: distribute 2,3,3,4,7,8,9 over 3 bins
Silly-heuristic
Bin 1: 9+3=12
Bin 2: 8+3=11
Bin 3: 7+4+2=13
However zero variance is attainable in this example:
Bin 1: 9+3=12
Bin 2: 8+4=12
Bin 3: 7+3+2=12
.
- References:
- Packing problem
- From: news.skynet.be
- Packing problem
- Prev by Date: Re: nth term of a Series when d is not constant
- Next by Date: Re: nth term of a Series when d is not constant
- Previous by thread: Re: Packing problem
- Next by thread: Reflexive Vector Space
- Index(es):