caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Radu Grigore <radugrigore@gmail.com>
To: fa.caml@googlegroups.com
Cc: caml-list@inria.fr
Subject: Re: [Caml-list] Priority queues, reloaded
Date: Sat, 2 Jul 2011 15:42:07 -0700 (PDT)	[thread overview]
Message-ID: <43c8bcb0-e306-421f-a379-43dd4e7de4a0@glegroupsg2000goo.googlegroups.com> (raw)
In-Reply-To: <30775f1d-9cce-4c46-b7f0-62cdcf9e52cf@glegroupsg2000goo.googlegroups.com>

I asked:
> How is that [a leftist heap implementation] better than using Set?

and Andrew answered:
> Sets are not heaps ; they can't have multiple copies of the same element.

I asked my question in the context of Andrew's original query:
>  I'd love to hear about efficient-yet-easy/fast-to-implement options.

I meant to ask why would it be easier/faster to implement a leftist heap. Of course, if leftist heaps can do something that an implementation based on Set can't then that would be a problem. Also, if an implementation based on Set would be much slower or memory hungry, that would also be a problem. But it was already explained by others for Dijsktra at least the lack of Decrease-Key is not a problem, and (asymptotic) performance is not an issue.

You now bring up the question of multiplicity, which may be important if you try to, say, find the top K elements in a stream. Again, a Set based implementation is much simpler than the leftist heaps that were posted already. Here it is.

  module S = Set.Make (struct type t = int*int*int let compare = compare end)
  let uid = ref 0
  let push pq p x = incr uid; S.add (p, x, !uid) pq
  let pop pq = let (_,x,_) as k = S.min_elt pq in (x, S.remove k, pq)

With a time limit of three hours this is what I'd do. In a real program, I'd probably go for binomial heaps if imperative is OK, or maxiphobic heaps if persistence is important.


  parent reply	other threads:[~2011-07-02 22:42 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <fa.V8myB/rA6OKILQg+GW40f8c1BGo@ifi.uio.no>
2011-07-02 12:24 ` Radu Grigore
2011-07-02 19:05   ` Andrew
2011-07-02 22:42   ` Radu Grigore [this message]
2011-07-10 17:55     ` Jon Harrop
2011-07-09 18:45 james woodyatt
     [not found] ` <14B0DF03-EF83-4568-AB34-6B51BCE4B574@recoil.org>
2011-07-09 18:56   ` james woodyatt
     [not found] <848371343.3424870.1309454037170.JavaMail.root@zmbs3.inria.fr>
2011-06-30 18:03 ` Daniel de Rauglaudre
     [not found] <sfid-j-20110630-131704-+2.76-1@multi.osbf.lua>
2011-06-30 17:13 ` Andrew
2011-06-30 17:26   ` Gabriel Scherer
2011-06-30 18:14     ` Jean-Christophe Filliâtre
2011-06-30 18:36     ` Jean-Christophe Filliâtre
2011-07-09  9:02       ` Jon Harrop
2011-07-09 19:22         ` Jean-Christophe Filliâtre
2011-07-10 18:04           ` Jon Harrop
2011-06-30 19:13     ` Andrew
2011-06-30 22:17     ` Wojciech Meyer
2011-07-02  1:49   ` Norman Ramsey
2011-07-09  9:05   ` Jon Harrop

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=43c8bcb0-e306-421f-a379-43dd4e7de4a0@glegroupsg2000goo.googlegroups.com \
    --to=radugrigore@gmail.com \
    --cc=caml-list@inria.fr \
    --cc=fa.caml@googlegroups.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).