I implemented the non-interval based simulation of modes so that ICP can decide which sub-interval to search first. It can be activated with the --ode_sim_heuristic flag.
I also fixed a bug in the CAPD formulation that results from double negation of ODEs. For example, in the backward integration of dx/dt = -y it would attempt dx/dt = --y, which CAPD does not like. It is now dx/dt = y in this case.
I implemented the non-interval based simulation of modes so that ICP can decide which sub-interval to search first. It can be activated with the --ode_sim_heuristic flag.
I also fixed a bug in the CAPD formulation that results from double negation of ODEs. For example, in the backward integration of dx/dt = -y it would attempt dx/dt = --y, which CAPD does not like. It is now dx/dt = y in this case.