Problem solving ode

[Alelafar]

I tried to use VAL in a very simple case of a durative action with continuous change.

We have a robot and one action with fixed duration. We have also a numeric function, 'battery_level', that decrease during the action continuously according the ode

     d(battery_level)/dt = - 0.01* battery_level 

whose solution is a simple exponential. Here the code:

(:durative-action move
  :parameters ( ?l_from - location ?l_to - location)
  :duration (= ?duration 5)
  :condition (and 
               (at start (connected ?l_from ?l_to))
               (at start (robot_at ?l_from))
               (over all (<= 0 battery_level))
             )
  :effect (and
              (at start (not (robot_at ?l_from)))
              (at end (robot_at ?l_to))
              (decrease (battery_level) (* #t (* 0.01 battery_level) ))
          )
 )

As you see I put as condition the invariant

(over all (>=  battery_level  0) ) 

and the validation is fine.
Unfortunately, if I just change the previous condition with the equivalent

(over all (<=  0  battery_level)) 

VAL fails. Here the error message:

Plan failed because of unsatisfied precondition in:
Invariant for (move l0 l1)

Plan failed to execute

Plan Repair Advice:

Invariant for (move l0 l1) has its condition unsatisfied between time 0 to 5, the condition is satisfied on [ 0 , 5 ]. 
(Invariant -100e^{ - 0.01t} >= 0 for t in (0,5) is only satisfied on [ 0 , 5 ])

The problem is in the first minus sign in the solution of the ode that should not be there and that fails the invariant condition. Why does it happen?

I leave the domain, problem and plan files.

ODE_problem.txt
ODE_domain.txt
ODE_plan.txt

[DerekLong101]

Hi Alelafar Thanks for this report. I will take a look at it. Cheers Derek

…

________________________________ From: Alelafar ***@***.***> Sent: 25 January 2024 16:47 To: KCL-Planning/VAL ***@***.***> Cc: Subscribed ***@***.***> Subject: [KCL-Planning/VAL] Problem solving ode (Issue #62) I tried to use VAL in a very simple case of a durative action with continuous change. We have a robot and one action with fixed duration. We have also a numeric function, 'battery_level', that decrease during the action continuously according the ode d(battery_level)/dt = - 0.01* battery_level whose solution is a simple exponential. Here the code: (:durative-action move :parameters ( ?l_from - location ?l_to - location) :duration (= ?duration 5) :condition (and (at start (connected ?l_from ?l_to)) (at start (robot_at ?l_from)) (over all (<= 0 battery_level)) ) :effect (and (at start (not (robot_at ?l_from))) (at end (robot_at ?l_to)) (decrease (battery_level) (* #t (* 0.01 battery_level) )) ) ) As you see I put as condition the invariant (over all (>= battery_level 0) ) and the validation is fine. Unfortunately, if I just change the previous condition with the equivalent (over all (<= 0 battery_level)) VAL fails. Here the error message: Plan failed because of unsatisfied precondition in: Invariant for (move l0 l1) Plan failed to execute Plan Repair Advice: Invariant for (move l0 l1) has its condition unsatisfied between time 0 to 5, the condition is satisfied on [ 0 , 5 ]. (Invariant -100e^{ - 0.01t} >= 0 for t in (0,5) is only satisfied on [ 0 , 5 ]) The problem is in the first minus sign in the solution of the ode that should not be there and that fails the invariant condition. Why does it happen? I leave the domain, problem and plan files. ODE_problem.txt<https://github.com/KCL-Planning/VAL/files/14055136/ODE_problem.txt> ODE_domain.txt<https://github.com/KCL-Planning/VAL/files/14055141/ODE_domain.txt> ODE_plan.txt<https://github.com/KCL-Planning/VAL/files/14055144/ODE_plan.txt> — Reply to this email directly, view it on GitHub<#62>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AB2KOVHI43AANPQ55TOD3DDYQKECRAVCNFSM6AAAAABCK23AVKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEYDANZYHAYDQNI>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[DerekLong101]

Can you just check for me what happens if you use (<= 0 (battery-level)) (ie with parentheses around the numeric fluent)? Thanks Derek

…

________________________________ From: Alelafar ***@***.***> Sent: 25 January 2024 16:47 To: KCL-Planning/VAL ***@***.***> Cc: Subscribed ***@***.***> Subject: [KCL-Planning/VAL] Problem solving ode (Issue #62) I tried to use VAL in a very simple case of a durative action with continuous change. We have a robot and one action with fixed duration. We have also a numeric function, 'battery_level', that decrease during the action continuously according the ode d(battery_level)/dt = - 0.01* battery_level whose solution is a simple exponential. Here the code: (:durative-action move :parameters ( ?l_from - location ?l_to - location) :duration (= ?duration 5) :condition (and (at start (connected ?l_from ?l_to)) (at start (robot_at ?l_from)) (over all (<= 0 battery_level)) ) :effect (and (at start (not (robot_at ?l_from))) (at end (robot_at ?l_to)) (decrease (battery_level) (* #t (* 0.01 battery_level) )) ) ) As you see I put as condition the invariant (over all (>= battery_level 0) ) and the validation is fine. Unfortunately, if I just change the previous condition with the equivalent (over all (<= 0 battery_level)) VAL fails. Here the error message: Plan failed because of unsatisfied precondition in: Invariant for (move l0 l1) Plan failed to execute Plan Repair Advice: Invariant for (move l0 l1) has its condition unsatisfied between time 0 to 5, the condition is satisfied on [ 0 , 5 ]. (Invariant -100e^{ - 0.01t} >= 0 for t in (0,5) is only satisfied on [ 0 , 5 ]) The problem is in the first minus sign in the solution of the ode that should not be there and that fails the invariant condition. Why does it happen? I leave the domain, problem and plan files. ODE_problem.txt<https://github.com/KCL-Planning/VAL/files/14055136/ODE_problem.txt> ODE_domain.txt<https://github.com/KCL-Planning/VAL/files/14055141/ODE_domain.txt> ODE_plan.txt<https://github.com/KCL-Planning/VAL/files/14055144/ODE_plan.txt> — Reply to this email directly, view it on GitHub<#62>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AB2KOVHI43AANPQ55TOD3DDYQKECRAVCNFSM6AAAAABCK23AVKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGEYDANZYHAYDQNI>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[Alelafar]

Yes, I tried but nothing changes. There is always the same error message.

Thanks,
Alessandro