AsyncTask creation life cycle is actually managed by a threadpool hide behind in all API level, this threadpool becomes accessible since API level 11, however, this does not means we can't create multiple AsyncTask instance before API lever 11, check out my SO question here for more details.
Come back to your topic, you can always pre-analyse your problem set and split/scale them into sub-set, for example, total 100 rss feed can be split into 5 chunks (with each perform 20 RSS feed in sequence), then feed these pre-processed sub-sets into 5 AsyncTask:
ArrayList<ProblemSet> problemSets = splitBigProblemSet();for (ProblemSet problemSet : problemSets) { new ProblemResolverAsynTask().execute(problemSet);}This will gives you 5 AsyncTask running asynchronously and probably get 5 times faster than using 1 AysncTask download 100 RSS feed. Note that it is also possible create AsyncTask per each RSS feed, which IMO is a very bad practice, as if your problem set is big, it will eat up you memory and kill system quickly.
Hope this help.
