Speed up logical operations

I am working on a problem: an array of sub-systems(over hundreds) , each has two states [0 1].

The pre-defined probability transition between are p1(0->1) and p2 (1->0)

The whole system is evolving with time, I am trying to use the logical operations to update the system state:

given an array of random number generated (pn)

SubSysState(SubSysState==0)=(p1(SubSysState==0)>pn(SubSysState==0))
SubSysState(SubSysState~=0)=(p2(SubSysState~=0)<pn(SubSysState~=0))

I did the 'Profiler' on the code, the above operations are lines where the most time was spent.

I am just wondering if there any other way to speed up it or anyone has better idea to do it?

Thanks a lot

Answer 1

[Sean de Wolski]

Compute the logical conditions only once:

SubSysStateeq0 = SubSysState==0;
SubSysStateneq0 = ~SubSysStateeq0;
SubSysState(SubSysStateeq0)=(p1(SubSysStateeq0)>pn(SubSysStateeq0))
SubSysState(SubSysStateneq0)=(p2(SubSysStateneq0)<pn(SubSysStateneq0))

[Feng]

Thank you, Sean

It did speed up a little. I should have thought about it.