1

I have model simple NOT gate in uppaal with propagational delay. But I am unable to prove one property. Screenshot of automata is attached. Xinp is input, xout is output. 3 is used as propagational delay here. 

P1: A[] (xinp && t>3 imply (xout==!xinp))

This property is working correctly , after 3 time unit, output becomes negation of input.

P2: A[] (xinp && t<3 imply (xout==!xinp))

This property should be unsatisfiable so that I can verify that not gate is working after the propagational delay and before 3 time unit output is not according to not gate. BUT this property is not unsatisfiable.

Thanks enter image description here

thor
  • 21,418
  • 31
  • 87
  • 173
Qurat
  • 64
  • 5
  • 1
    Double negatives are confusing in English: what do you mean by "property is not unsatisfiable"? The property is either satisfied or it is not. In this case if A[] property is not satisfied, then there is a counterexample (and Uppaal should provide one). Does xinp change over time? – mariusm Mar 30 '17 at 07:54
  • i mean property should be unsatisfiable, so that i can verify Not gate is not working before 3 time unit i-e output is inverse of input only after the 3 time unit. Counter example showing error in s00 and outgoing edge. xinp is defined as a boolean and not changing over time. – Qurat Mar 30 '17 at 08:11
  • as I understand your Uppaal counter example says that the property is indeed *not* satisfied (trivially) at s00. So I don't see any contradiction except ambiguous expression "not unsatisfiable" -- it must be either satisfied, or not satisfied. – mariusm Mar 31 '17 at 09:11

0 Answers0