[jira] [Commented] (GIRAPH-824) Concurrent Random Walks on Graphs

["George Valkanas (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/GIRAPH-824?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13889004#comment-13889004 ] 

George Valkanas commented on GIRAPH-824:
----------------------------------------

Quick question, before generating the patch:
Do you think it's best to have the Concurrent Random Walks as a separate set of classes, or should it replace the existing one (i.e., replace the single RW with the general case)?

I'm in favor of the 1st alternative (for various reasons).

I can submit a patch, as soon as I have an answer.

Thanks!

> Concurrent Random Walks on Graphs
> ---------------------------------
>
>                 Key: GIRAPH-824
>                 URL: https://issues.apache.org/jira/browse/GIRAPH-824
>             Project: Giraph
>          Issue Type: New Feature
>          Components: examples
>    Affects Versions: 1.1.0
>            Reporter: George Valkanas
>              Labels: examples, graph, multiple, randomwalk
>         Attachments: ConcurrentPageRankComputation.java, ConcurrentRandomWalksComputation.java, ConcurrentRandomWalksVertexMasterCompute.java, ConcurrentRandomWalksWithRestartComputation.java, ConcurrentRandomWalksWorkerContext.java
>
>
> In certain cases, running multiple random walks on the same graph may be desired (e.g., the centerpiece algorithm). The current implementation(s) for random walks on graphs are designed for a single random walk. Therefore, running multiple random walks can be done sequentially or by spawning multiple independent random walks, which can be inefficient and harder to manage within GIRAPH.
> The idea is to extend the basic Random Walk examples so that the vertices are aware that they participate in multiple random walks at the same time.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)