Future design ideas

[javier-romero]

Hi,

Here are some of my ideas related to the future design of telingo:

1. Apply a lazy grounding approach where the temporal program is only grounded for two consecutive time points.
2. Apply a method for temporal nogood learning where nogoods learnt over some specific time points can be generalized to all time points.
3. Develop the previous extensions to consider a more general case where time is not the only "parameter". For example, in asprilo we could consider also the robots as parameters, and then we would only have to ground the program for one robot and two consecutive time points, and the learnt nogoods could be genrealized to all time points and all robots.

Maybe the point 3 won't be part of telingo, but I think we should take it into account when we decide about the design of telingo.

Chao!
j