Since we have to compute multiple reach/evolve sets for DI (due to the more accurate being non rigorous with respect to the reach set, and due to looking for the best approximation), it makes sense to parallelise it. The advantage will not be very significant in most cases since we already optimise along time to use only the most effective approximation.
Since we have to compute multiple reach/evolve sets for DI (due to the more accurate being non rigorous with respect to the reach set, and due to looking for the best approximation), it makes sense to parallelise it. The advantage will not be very significant in most cases since we already optimise along time to use only the most effective approximation.