Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

I Markov property: compatible with momentum?

  1. Jun 6, 2016 #1
    Say I'm simulating the movements of two interacting atoms. Could this system have the markov property (the future positions of the atoms depend only on the current position, not the past)?
    What's got me on the fence are the atoms' momentums: it's a property of the present state (at time t, the momentum is p), but its value depends on past states. So does anyone know if a system like this could have the markov property?
  2. jcsd
  3. Jun 6, 2016 #2


    User Avatar

    Staff: Mentor

    You misunderstood what Markov is about. Of course, the state of any system at time t0 depends on what happened at times t < t0. But in a Markovian process, there is no memory of how the state at time t0 was reached. All that is important is that the state at time t0 completely describes the system (and its future evolution).

    In the case of two atoms governed by a known Hamiltonian, given the wave function ψ(t0) of the two-atom system, one can calculate ψ(t) for any t > t0. No need to know the wave function for any time < t0.
    Last edited: Jun 6, 2016
  4. Jun 7, 2016 #3

    Stephen Tashi

    User Avatar
    Science Advisor

    It isn't meaningful to ask whether a physical system has the markov property until you say what variables you will choose to define a "state". A physical system may have the markov property when one definition of "state" is used and not have it when a different definition is used.

    Are you considering actual atoms - or just thinking about "point particles" ?

    You only mention "position" as the variables involved in your definition "state" ? Did you intend to omit variables like mass, velocity and acceleration ?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted